Some Realism about Pluralism: Legal Realist Approaches to the First Amendment– Part II

Originally published in 1990 Duke L.J. 375
Copyright 1998 by Jack M. Balkin. All Rights Reserved.

by J.M. Balkin*




The above analysis of public forum law argues that our freedom to speak depends upon access to particular forms of property. More generally, access is determined by control of what we might call the “means of communication,”(68) which include access to various places to speak (public and private real property) and access to various ways of speaking (for example, publishing houses, television stations, and other communication media). One can have control over the means of communication because one owns them outright, or because one has purchased access from those who do own them, or because the law requires the owner to give access.

As noted above, Justice Roberts’ solution to the problem of access to the means of communication – the creation of a first amendment easement in streets and parks – proceeded from the fiction that expressive activity was a traditional use of streets and parks. Yet this form of justification had several unfortunate consequences. By this fiction, Justice Roberts seemed to establish that the government, like an owner of private property, had the right to exclude persons from the expressive use of its property, except where it traditionally had suffered their presence. Moreover, by emphasizing that access to streets and parks was traditional (an unjustified assumption in any case – most municipalities up to that time wanted nothing less than protests in the streets),(69) Roberts did two things. First, he tied the question of access to the question whether a particular form of access was of the type that the state had previously allowed. Second, he tied the question of access to traditional technologies of communication – that is, protesting in streets and parks, handing out leaflets, and so forth. However, Roberts wrote his opinion as technologies of communication were rapidly changing. It soon would become clear that these traditional technologies of communication were as efficacious in a world flooded by the communications of mass media as a blacksmith’s forge in an era of mass industrialization.

I have just spoken of “technologies of communication,” and I believe that this concept is crucial to a realist analysis of first amendment law.(70) Although we normally think of liberty as freedom from restraint and therefore as wholly unrelated to the existing state of technological achievement, in fact many human liberties are dependent on forms of technology for their exercise. Where the exercise of a liberty depends upon technology, access to that technology largely determines the substantive liberty of the actor. Sometimes this is because the liberty in question cannot be enjoyed in any form without a given level of technology. More commonly, however, it is because liberties are always in conflict. Access to widely different levels of technology by persons who seek to exercise competing liberties may place some actors at a very significant disadvantage with respect to others, and thus result in an effective denial of their liberty.

In the paradigmatic situation of the speaker on the soapbox, technology appears to play no part in the exercise of expressive liberty. One simply speaks, and one’s voice is heard by others. These others are either convinced or not convinced, and further speech acts ensue, all using the “natural” tools of the human voice. In contrast, we recognize the pervasive role that technology plays in the production of wealth and economic value, at least once civilization has progressed past the point where wealth is created through brute labor alone. Indeed, as soon as tools are invented, the production of wealth depends upon technology. In present day America, reliance on technology and capital investment in the production of wealth is taken for granted. Wealth is not created simply through individual effort – rather we need technology to transform our labor into goods and services. Moreover, the type of labor that must be input is determined in large part by the requirements of available technology. Strength is required for some jobs, finesse for others. Physical stamina may be essential to the manual laborer, whereas a different sort of stamina may be needed by the bond trader or arbitrageur.

Yet many of the things that can be said of production of goods and services are true of communication and the production of information. Just as the power to create wealth relies to a large extent upon the existing technologies of the time, so does the power to communicate with and influence people. Of course some people are influential or good at communicating because they are better orators, painters, musicians, or writers than others. But the same is true of wealth creation. Some people are more skillful at using the existing technologies that produce wealth. Talents for effective communication do not simply preexist society, but also depend upon and adjust to, the nature of the technologies of the time. A great orator in Periclean Athens may have needed only a loud voice and a fluid style; in the era of Ronald Reagan, he must look natural in makeup and have some degree of facility with a Teleprompter.

The dependence of effective exercise of rights upon technology is a major source of the divergence between formal and substantive liberty, and this is as true of liberty of speech as it is for the liberty of contract. If ownership, control, or other access to the means of production are essential to wealth creation, then ownership, control, or access to the means of communication are essential to effective communication. To put it bluntly, the more property one has, the greater one’s ability to compete in the marketplace of ideas, just as in the ordinary marketplace. Similarly, to the extent that one does not own the means of communication, one must bargain with others to obtain access. We still retain a romantic vision of the great thinker who changes history through the strength of her ideas and the power of her charisma. But increasingly, technology swamps such “native ability.” The president of a large corporation may not be as fluent a speaker as William Jennings Bryant or Jesse Jackson. But what does she care if she can hire the best advertising agency to formulate her message and the most attractive actor to recite it?(71)


This brings me back to Justice Roberts and the doctrine of the public forum. In his solution to the problem of access, Roberts chose a method that in the short run appeared not much less protective of access than a tax on or easement to private property. After all, protests on a neighbor’s land next to a public park are probably no more effective than protests in the park itself. Thus, there was no compelling reason to create easements to private property when access to public property would do just as well. Nevertheless, as new technologies of communication out-stripped older forms in terms of effectiveness, there was greater and greater significance to the difference between access to traditional forms of communication (available under public forum law) and access to newer forms that could only be purchased from private sources. There may well have been as many streets and parks as there were parcels of private land when Roberts wrote. But it soon would become clear that this parity was not preserved for other technologies. For each private newspaper, or television or radio station, there was not a corresponding government-owned forum open to all on a first come, first served basis.

Professor Fiss captures the essence of the problem when he argues that speech, like other resources, exists under conditions of scarcity.(72) This scarcity is of at least two types. First, effective communication (and that is, on the whole, the only sort of communication most people are interested in) costs money. Like all other desirable things, technologies of communication are scarce, and this has been true even when those technologies were primitive. If one person has a pleasant voice, it costs money to hire that voice as an advocate. If another person has a plot of land useful for organizing, the right to use that land costs money. If still another is the only literate person in a small village and therefore can transfer thoughts to paper and decipher them, this is also a technology of communication that costs money.(73)


Improved technology does not change this feature of scarcity. Even after printing presses are invented, some own them, while others do not. Paper costs money, as do typefaces, ink, and newspaper buildings. Educating people to read and write (which we tend to forget is necessary to much effective communication) involves an enormous investment of resources, so great in fact that it is often treated as a public good and delegated to the government. I suspect that if we opened up the air waves to competitive bidding – as some have suggested(74) – at the end of the bidding there would be no more space left, given a particular level of technological development. Where the technology that constitutes the dominant means of communication is inexpensive, we can expect that the problem of scarcity, and hence the distribution of property rights, will have somewhat less effect on speech. Yet as society changes and effective means of communication become increasingly expensive, the right to effective speech becomes increasingly linked to the distribution of property.

Effective communication is scarce not only in the sense that technologies of communication are limited. Communication is scarce also in the sense that there is only so much available audience time to go around. Although newer technologies like the mass media can reach more people more quickly, they still do not eliminate this second type of scarcity. Indeed, mass communication only increases the competition for audience attention. Simply put, if thirty percent of the American public is watching “Roseanne” on Tuesday nights, we can rest assured that they will not be listening to a speaker in the park criticizing U.S. foreign policy or reading a book on the history of American religion. In an earlier age in which one could reach only a limited audience with one’s voice, it may have seemed that there was a plenitude of listeners, and audience scarcity was not a real phenomenon.(75) With the advent of mass media, however, we see all the more urgently that speech rights can come into conflict not only with the property rights of others, but also with the speech rights of others.

Of course this problem as well always existed in the abstract. To have a liberty interest (or “privilege” in the Hohfeldian sense(76)) meant that one could not be prevented by the state from engaging in certain behavior. It did not mean, however, that one could not be prevented from exercising that liberty because other private parties had soaked up all of the available resources for its effective exercise. One might have the freedom to park one’s car in any space in a municipal parking lot, but that freedom is meaningless if all of the spaces are occupied by other private parties.

Thus my speech and your speech are always potentially in competition with each other. This is due partly to audience scarcity and technological scarcity. It is due also to the fact that one of the ways we exercise our liberty of speech is by not speaking, or rather by not having certain ideas or beliefs associated with us. For example, if we give protesters access to shopping malls, then the owner of the mall’s freedom not to speak is endangered. Some people like to think that this is not a true abridgement of the freedom to speak because the shopping center owner can simply post a sign saying that she does not agree with the speech being made. I think this explanation simply defines the problem out of existence. You should try explaining this theory to a black entrepreneur who has to let the Klan march through her shiny new food court and see how receptive she is to the idea.

In like fashion, if we give Vietnam war opponents access to CBS’s technology, we are diminishing CBS’s right to speak, not to mention imposing an opportunity cost on it for the air time it could have sold for other purposes. In short, we cannot guarantee freedom of speech for everyone anymore than we could have vindicated everyone’s conflicting freedom of association rights in Brown v. Board of Education.(77)

Let me summarize the argument so far. Modern public forum doctrine has obscured a fact about the right of free speech: Speech rights depend upon access to communicative technologies, which are forms of private property. The effective exercise of speech rights thus both depends upon one’s own property rights, and is potentially in conflict with the property rights of others. Moreover, because of the relationship between speech rights and technology, speech rights are potentially in conflict with other speech rights.

It is not difficult to see the relevance of legal realism here. The legal realists taught us to look beyond the division of the world into public and private and recognize that the state is largely responsible for forms of private power that interfere with the effective exercise of private rights. Thus, for example, Robert Hale argued that a certain degree of coercion inheres in all economic transactions, and this coercion is due to the constellation of property and contract rights provided by the state. One has to contract with others to purchase food because they have property rights in food that the state will enforce.(78)

Nothing, however, has changed when we move from bread to broadcasting. One can have control over the means of communication because of ownership, purchase of access, or access enforced by law. Yet following the legal realists, we might note that the third situation actually includes the first two. Ownership of the means of communication and the right to sell or to refuse to sell access are state-enforced guarantees and denials of access. Whenever the state enforces an advertising contract, allows a person to own and run a newspaper, or grants a broadcast license, it is sanctioning a grant of access to the means of communication. Whenever the state ejects a protester from private property, or enforces the right of CBS to refuse to broadcast the views of Vietnam War opponents, it is sanctioning a denial of access to the means of communication. If freedom of contract is a state enforced monopoly in the use and disposition of the means of production, then freedom of speech can be reinterpreted as a state enforced monopoly in the use and disposition of the means of communication.

This type of analysis holds true even in the simple case of individual speech. The right to speak is the state’s sanction to use one’s voice to convince others (without fear of direct state punishment), as well as the right not to be made a mouthpiece to shout the slogans of others. Note, however, that the state simultaneously guarantees the right of private parties to harm a speaker’s interests because of the content of one’s speech – for example, to refuse to associate with the speaker, to refuse to sell air time or newspaper space to the speaker, or to boycott the speaker’s business. Private forms of speech control thus owe their efficacy in part to the existing private rights of social and economic power guaranteed by the state. The right to boycott a business owned by a racist would mean nothing if one did not have the right to purchase one’s goods elsewhere. And, as we have already seen, the right to deny access to broadcast and print media exists only because of the state’s rules of property and contract.(79)

If freedom of speech is a state-granted monopoly in the use and disposition of the means of communication, it becomes increasingly difficult to see the liberty of speech as merely a grant of formal equal liberty to speak, unrelated to issues of substantive equality. It becomes problematic to claim that the state has not exercised a substantive choice when a William Loeb or Ruppert Murdoch can reach a large number of people, and persons with opposite but equally extreme views can reach very few.(80)


Of course, one might respond that the state is involved in these cases – that it does intervene and that its intervention is not value-free – but that the value that it imposes is one of neutrality. However, if the state is always implicated in access to speech, if the state’s contract and property laws always help determine one’s freedom of speech because they determine access to the means of communication, then it is difficult to argue that the state is neutral when some persons have vastly greater access to the means of communication than others because of vastly more economic power. The state is no more neutral in these decisions than it is when it enforces or chooses not to enforce contracts of adhesion or contracts with unconscionable terms.

One might object that whatever one’s feelings about the assumption that formal equality does not involve state interference with liberty, this theory is written into our Constitution because the first amendment says that Congress shall make no law abridging the freedom of speech. However, this argument proves too much. Under a theory of formal equality, the government would have no obligation to provide any public forums at all. Moreover, if one accepts the force of the legal realist critique as applied to contract, the rules of property themselves affect and therefore may potentially “abridge” the freedom of speech. It all depends upon what you mean by an “abridgement,” and we are back once again to the issue of substantive liberty.

I realize that these conclusions seem to discount the value of intellectually safe harbors like formal equality and content neutrality. But my point is that once the legal realist critique of economic liberty is applied to the first amendment, these safe harbors can no longer seem quite as safe as they did before. Certainly I do not pretend to have complete solutions to the problem of access. I am simply asking that we abandon belief in a rigid division of public and private spheres in the realm of communication and information that we jettisoned long ago in the areas of property and economic exchange. A libertarian conception of free speech has served us well in the past, but like all conceptions, it can and eventually must run out of steam and degenerate into a sterile conception that will hinder progressive reform rather than aid it.

What I have said suggests that redistributive legislation might be a good means of enhancing the substantive liberty of speech (and many other liberties as well, one might add). It also suggests that governmental investment in the modern technological equivalents of traditional public forums – for example, radio and television – would help ameliorate the situation. Whether such expenditures are required constitutionally by the first amendment is a more complicated question. Yet if (as I believe to be the case) the first amendment requires the government to create at least some public forums that provide effective means of communication, I believe the answer to that question must be yes.

The key word in the last paragraph, it seems to me, is “effective.” Just as I accept the legal realist argument that questions of economic duress and substantive unconscionability are matters of degree in the determination of economic liberty, I also believe that the question of effective access to the means of communication is always a matter of degree. This does not mean that once having weighed the relevant factors, we should not adopt rules that approximate our concerns but have the virtue of being relatively easy to apply. The point I am making is true of every affirmative liberty (such as education), and it is especially true of every negative liberty that turns out to be an affirmative liberty (like speech). Where affirmative liberties are at stake, the most that courts can do is define a range of alternatives for the political branches to pick from, or direct the political branches to propose their own alternatives and then accept them if they appear reasonably calculated to succeed. In other words, the effective protection of affirmative liberties requires considerably more judicial restraint than the protection of negative liberties.

At this point a few words about campaign finance are in order. The arguments I have just given, which tie the liberty of free speech to communicative technology and thus to property rights, may seem to concede too much to the “money is speech” position used in the past to thwart campaign finance reform. But I think this concession (if it is a concession) is intellectually necessary for a legal realist analysis to proceed. If control of the means of communication is necessary for effective speech, and if such control requires property, then speech rights and property rights are intimately related – property is what gives one access to the means of communication. The entire argument has been based upon the assumption that one of the best ways to shut someone up is to impoverish them. If property and speech rights are intimately related, that is all the more reason to regulate property used to influence the outcome of political campaigns.

I suspect that the slogan “money is not speech” is attractive because it appeals to a certain humanistic vision – that there is something quite different between the situation of a lone individual expressing her views and the purchase of hired mouths using hired expressions created by hired minds to saturate the airwaves with ideological drivel. Yet in one sense, this humanistic vision really turns upon a set of unstated egalitarian assumptions about economic and social power. Certainly we would have no objection to a person with a speech impediment hiring someone to do her talking for her; that is because we think that, under these circumstances, it is fair for such a person to boost her communicative powers. Modern political campaigns seem a far cry from this example because of the massive amounts of economic power expended to get the message across. I think we should isolate the egalitarian assumptions implicit in the “money is not speech” position and put them to their best use – the justification of campaign finance reforms on the grounds that gross inequalities of economic power destroy the integrity of the political process.


My conclusion, then, is that campaign finance reforms may be constitutional not because money is not speech, but because in a very important sense it is. The government is responsible for inequalities in access to the means of communication because it has created the system of property rights that makes such inequalities possible. Therefore, it is not only wrong but also incoherent for opponents of campaign finance reform to contend that the government should not regulate access to the political process. Government already regulates access to the political process – the first amendment simply demands that it do so fairly. At the very least the first amendment should not act as a barrier to attempts to ensure that the process works equitably. Thus, the advantage of a legal realist analysis over more traditional approaches is that it allows one to hoist the opponents of campaign finance reform by their own petards.



We can generalize our previous discussion of public forum law in the following way: The right of free speech does not consist merely in protecting citizens from being harmed by the government. It also includes the government’s grant of power to private citizens to harm others through the exercise of their right to speak. It is a statement by the government that a particular exercise of power is permissible, and that the other party has no right to prevent the exercise of that power. Moreover, protecting freedom of speech also involves protecting the freedom not to speak, that is, protecting a person’s right not to be associated with a particular type of speech, or the person’s right to deny access to a particular means of communication that she controls. Rights to speak and not to speak, to grant access and to deny access, are thus delegations of public power; public rights against the government are also private rights against others.

Whenever the government grants private parties the right to withhold benefits to others, issues of coercion arise. We have already seen that a system of property rights “coerces” persons to bargain for access to the means of communication. But speakers are not the only persons who suffer coercion – the recipient of the message also can be coerced into listening. In the usual libertarian discussion of free speech, the problem of audience coercion does not exist or is relatively insignificant. The listening party is free to listen, avert her ears, or engage in counterspeech of her own. To be sure, there are exceptions. Some regulation of speech is permitted where there is a captive audience, because in that case the listening party is forced to listen against her will.(81) The libertarian vision thus rests upon an important distinction between exceptional situations in which there is a captive audience and the usual or normal case of speech in which these problems do not exist to any significant degree. Put even more generally, the libertarian vision rests upon a distinction between communication in conditions of free will and communication under duress.

Once we recharacterize the situation in this way, however, it is clear that the problem of the captive audience is much like the problem of access. There is simply no bright line test to tell us whether a situation of speech involves coercion or not.(82) Although we might wish for a world divided into audiences who had the free choice to avoid the speaker and those who were forced to listen under duress, the world is not so constructed. It is all a matter of degree. Indeed, we can go further and note that just as all contractual situations can be reconceptualized as involving a form of economic coercion, all speech situations involve different degrees of coercion as well.

This claim must surely seem odd at first. Yet it is a simple application of an argument made by the legal realists long ago in the context of economic liberty. Here we must turn again to the work of Robert Hale. Hale pointed out that the reason why we contract with other people is because they have property rights. The coercion inherent in market transactions consists precisely in the fact that others can refuse to deal with us or give us things we want unless we pay them for the privilege. Indeed, if we try to take something from them without contracting, they can invoke the power of the state to punish us.(83) Hale’s point was simply that you can coerce someone to do something when you have rights and can threaten to exercise them. Sometimes this coercion is not at all unpleasant, and we hardly notice it as such. In other cases, such as contracts of adhesion and cases of economic duress, our subjective experience is quite different. Nevertheless, Hale argued, coercion is simply the flip side of a guarantee of free choice to deny benefits to others. Coercion has no necessarily pejorative connotations; it is merely the by-product of a system in which private rights are protected by government sanction.(84)

For this reason, Hale argued, one should not assume that the existing regime of contract rights enforces only bargains entered into without coercion. The background allocation of property and contract rights sets the ground rules for how parties will be legally permitted to coerce each other. The appropriate question to ask is how much coercion the law will allow. If we have a classical theory of consideration and no doctrine of substantive unconscionability, then the coercive power produced by the exercise of private rights will run in one direction. If we substitute doctrines of detrimental reliance and strong notions of unconscionability, then the balance of coercive power will shift to other parties. In neither case, however, will we produce a system that respects only the free will of the parties and does not involve forms of coercion.

Indeed, one can make an even more general argument about free will and coercion. Free choice is an intellectual construct that occupies the semantic space that has not been assigned to the concept of coercion.(85) Nevertheless, because guarantees of private choice also produce opportunities for coercion, these two concepts exist in a relation of mutual dependence and differentiation. What we call freely chosen action is always circumscribed within a set of limitations on action. These limitations construct the contours and boundaries of what we call a person’s free choice. In most cases, it is perfectly reasonable to speak of a person who is limited by circumstances as nevertheless acting or choosing freely. The problem comes when we move to issues of justification. To justify existing limitations on action or choice based on the fact that one is not acting under duress but instead has free choice – which means only that one is choosing within the context of those limitations – is ultimately a circular argument.

This general point applies to limitations produced by rules of law. We say that actors within a system of law have the freedom to choose how they will act. But this freedom is circumscribed and defined by the set of limitations that the law imposes on the actor, as well as the powers of coercion granted to other private parties by the law. Thus, what we call free choice is not something that preexists the legal regime, and which the legal regime merely attempts to vindicate. Rather, free choice (and its opposite, coercion or duress) are constructed by the existing regime of legal rules. This leads to a problem of circularity like that described above. It will do no good, for example, to say that a contract with grossly unfair terms is just because the parties agreed to it through an exercise of their free will. The problem is that the free will of the weaker party is defined and constructed by what types of actions are available, given the existing system of rules of contract and property. It may be true that the weaker party chooses the unconscionable terms, but that is because the rules of property and contract do not allow her to force the other party to offer better terms. Thus a system of rules circumscribing the scope of one’s choices in economic bargains cannot be justified on the grounds that people acting within the system of rules freely choose what they think best for themselves given the legally available alternatives. For virtually any system of legal rules could be justified in this way.(86)


The same arguments about free will and coercion in the economic marketplace apply to the problem of coercion in the marketplace of ideas – that is, the problem of the captive audience. We feel sorry for the captive audience because we believe that persons who listen in such circumstances are listening against their will, lacking any real alternatives. In contrast, we note that the person who is offended by speech in other settings can, by an act of will, avert her eyes, escape the speaker, or stay and argue back.(87) If she stays and is offended or injured by the experience, her injury is a product of her own willed choice, and her offense, in some sense, is her own fault. My point is not to deny the value of this common sense way of looking at things. Rather, I want to emphasize that no less than in the case of contractual relations, what we call an exercise of “choice” and what we label a “captive audience” or the product of “duress” is the result of a background set of rights, which include not only property rights but also the right of free speech itself. In other words, Hale’s analysis of freedom and coercion, so admired by left scholars in discussions about labor legislation and welfare rights, must also be reckoned with in first amendment law, which also relies upon similar concepts of “choice” and “duress.”

Let us take, for example, the case of the communication you are presently reading. This poses few problems of unjust coercion between author and reader. You have chosen to read this Article. You can pick it up or put it down, scrawl nasty comments on the margins of the paper, and so on. If you are offended by what I am saying, nothing forces you to read further. You are exercising free choice. My point, however, is that even if you see your choice as free, it is also a choice made within a preexisting set of property and speech rights. The relatively non-coercive nature of this communicative transaction derives from the assumption that nothing substantial in your life (whether it be retaining your job, advancing your career, or impressing someone else) turns upon your reading or not reading this Article. On the other hand, if you wished to become a lawyer, and if everyone who graduated from law school was required to recite the contents of this Article by heart or produce a detailed analysis of its arguments (here I indulge in a law professor’s dream), you undoubtedly would think your freedom to refuse to read this Article was significantly diminished. You might retreat to the position that your will was unencumbered because you still had the choice, after all, to become something other than a lawyer. But this argument is strangely reminiscent of the Lochner-era argument that bakers are not in any sense forced to work more than sixty hours a week by their employers because they are in no sense forced to become bakers.

Indeed, once we understand that what we call “will” or “choice” is actually the product of personal predilections exercised within a constellation of governmental regulations and private expectations (which in turn are enforced or curtailed by other governmental regulations), it does not seem so odd to say that most law students are coerced into reading large volumes of material every day as a part of their training as lawyers, including specific documents such as the Constitution of the United States and the Uniform Commercial Code. The fact that substitutions are readily available in the marketplace of ideas (in the form of Gilbert’s and Emanuel’s) does not change the basic nature of the argument, other than to allow us to achieve a more precise definition of the boundaries of coercion and duress that inhere in our educational system.(88)


I want to leave these academic examples and turn to a slightly different set of problems. Suppose that we do not have a situation of an author and a reader, but rather a young black woman pursued across the quadrangle of a college campus by a group of young white males who hurl racial and sexual insults at her. In one sense, she is not a captive audience because she has (and is presently exercising) the free choice to avoid these men, dash into the nearest building, lock the door if a lock is available, and wait for them to grow tired of their sport and leave her in peace. Perhaps this exercise of will is all the first amendment demands to avoid the conclusion that she is a member of a captive audience. Perhaps, however, you will think that something more is required. In any case, we should note that her choice of how to respond to their speech is affected not only by numerical and physical disadvantages, but also by the fact that it is illegal for her to pull out a pistol and threaten her persecutors, or, what is equally important, to have the campus authorities discipline the students for engaging in these acts of speech. Put another way, the existing system of rights and obligations, including the free speech rights of her pursuers, affect her will, inhibit her will, indeed construct her will just as surely as the liberty of contract affected, inhibited, and constructed the will of the bakers in Lochner v. New York(89) or the employees in Coppage v. Kansas.(90)

If free speech doctrine is justified through concepts of will and free choice, and if will and free choice are constructed by the system of legal rights and obligations, including the rights of free speech, then there is an inherent problem of circularity. It is true that one always has the free choice to avert one’s eyes when one sees a naked buttock on the screen,(91) in the same sense that it is always true that the weaker party to an unconscionable bargain always has the free choice to walk away or to accept the unconscionable terms. But this “choice” tells us no more about the justness of the duress and coercion involved in the law of free speech than it does for the doctrines of laissez-faire capitalism.

I believe that if we assimilate the legal realist critique of contract into first amendment law, we will recognize that the concepts of coerced and non-coerced exposure to speech also exist in a relation of mutual dependence and differentiation. We will recognize that these terms do not preexist the system of first amendment law but rather are constructed by it, and that defenses of first amendment liberties in terms of freedom and coercion will prove ultimately as circular as those for freedom of contract did. This does not mean that most of current first amendment doctrine is wrong, or that we should start rounding up offensive speakers and throwing them in jail. I do suggest we recognize that our protection of free speech rights is protection of a certain type of coercion, of induced harm, and that we should be more sensitive to the existence of this coercion and this harm in specific and limited contexts – for example, direct face-to-face racial and sexual harassment.

To some extent, we already do recognize the problem of coercion through the fiction of “fighting words” – that is, words that by their nature are likely to incite an immediate breach of the peace. Nevertheless, the use of the “fighting words” doctrine to deal with face-to-face racial or sexual harassment is a very bad idea. It merely disguises and misrepresents what I believe is the real issue in these cases – that is, the harm forced upon an audience in an extreme and unfairly coercive situation. The problem with group harassment of the student in my previous example is not that as a result of their speech she is likely to fight back. The problem is that she is not going to fight back – that she will be intimidated and silenced by their heckling.(92) Both the rationale of the original “fighting words” decision, Chaplinsky v. New Hampshire,(93) and the later gloss provided by Brandenburg v. Ohio – that unprotected speech must be directed to produce imminent lawless action(94) – are ill-equipped to deal with cases of harassment for precisely this reason. The paradigmatic situation these cases are concerned with is the speaker who so angers her audience that they attack her, or so inspires them that they arise and revolt. These cases do not concern speakers who browbeat their opponents into silence. If in the first two cases we understand that no counter-speech will occur because of the imminence of violence, we also should understand that in the third case no counter-speech will occur because of the directness of the intimidation. Moreover, we should recognize that in the case of the inciting, as well as the harassing, speaker, judgments cannot be clear-cut but are always matters of degree. If there are problems of administrability in the latter case, there are also problems in the former case, which is already comprehended by current first amendment doctrine.

The most obvious place in which the Chaplinsky/Brandenburg doctrines of non-protected speech fail us is the case of sexual or racial harassment in the workplace. And here the conflict between the left-libertarian conception of free speech and the progressive agenda of guaranteeing racial and sexual equality is especially pronounced. The most rudimentary Hohfeldian analysis(95) demonstrates that to the extent we allow verbal conduct creating a hostile working atmosphere, we thereby refuse to protect persons from certain forms of private racial and sexual discrimination. Conversely, to the extent that mere words can give rise to liability for employment discrimination, intentional infliction of emotional distress, or other causes of action, we acknowledge that an employer or co-worker can be punished for making such statements.

This problem has not yet been squarely addressed in the courts.(96) I suspect that this is partly due to the limited nature of remedies for sexual harassment under Title VII. Under present law, Title VII(97) does not allow traditional legal damage remedies for sexual harassment. It permits injunctive relief to eliminate offending practices in “hostile environment” cases. Monetary relief is available only when there has been an actual or constructive discharge as a result of harassment, and even then Title VII only allows for the equitable remedies of reinstatement and back pay.(98) Nevertheless, awarding any sum of money for harms caused by speech surely raises first amendment concerns. Moreover, injunctive relief ordered against an employer to cease all harassing behavior in the future – including harassing speech – has many of the trappings of a prior restraint. Causes of action for intentional infliction of emotional distress, which allow for the whole panoply of traditional tort remedies, raise the conflict between first amendment values and egalitarian concerns in starker terms.(99) So too would the enactment of group libel laws, which presumably would operate outside the workplace.(100)

It is surprising that defense counsel have not regularly raised first amendment challenges to allegations of employer misconduct in the developing law of workplace harrassment. Aside from the limited remedies available under Title VII, I suspect that one reason for the strange paucity of first amendment defenses in sexual harassment cases is due to an unconscious form of categorization – that speech in the workplace is not considered speech in the same sense as political or expressive speech generally, but is thought to be utilitarian, pedestrian, and incidental to the performance of work. Of course, as soon as these categories are constructed, it is not difficult to break them down. And indeed, it is likely that very soon defense counsel will connect the first amendment attacks on campus regulation of racist speech with analogous situations in the workplace. We are likely to see increasing numbers of first amendment defenses raised in the years ahead. When and if litigants catch on to such possibilities, the clash between the left goal of egalitarianism and the libertarian theory of the first amendment will be felt with particular poignancy. One or the other has to give way, and I suspect that for many on the left it will be libertarian theory. This is yet another example of the phenomenon of ideological drift – the means by which the libertarian theory of the first amendment increasingly is turned to serve conservative social interests. In the not too distant future, then, we may well see defenders of racist and sexist employment practices join the Klan, cigarette manufacturers, and conservative PACs as the staunchest advocates of the principle of free speech.

The conflict identified here – between egalitarianism and the principle of content neutrality in the regulation of speech – is likely to manifest itself more and more frequently as time passes. It is important that we recognize these problems now and work towards understanding how to reconcile these competing interests before first amendment defenses of sexual harassment become routine.(101) The principle of remedying racial and sexual harassment in the workplace must have stronger protection than the present limits of innovation by defense counsel.

I think, in fact, that there are perfectly good ways to defend laws against racial and sexual harassment in the workplace from first amendment challenges. The question of sexual and racial harassment in the workplace ties in quite well with the analysis of captive audience doctrine presented above. Few audiences are more capitve than the average worker. It is true, in theory, that one does not have to be subjected to racist or sexist speech on the job – one can simply shield one’s eyes or ears, or failing that, one can decide not to show up for work anymore. But this will mean that one’s employment also will end. Because the will of employees is circumscribed by their need for employment and because employment is yoked together with the hostile work environment, traditional first amendment claims that more speech is better, and that one need not submit to distasteful speech, lose much of their force. Certainly, if employer-employee relations involve sufficient coercion that we can justify regulation in other contexts, then this coercion does not suddenly vanish when the issue is submission to racist or sexist speech.

If the workplace involves sufficient coercion to invoke the doctrine of the captive audience, then perhaps the first amendment problems I have identified above will not prove insurmountable. Indeed, I suggest that we might do well to shift the paradigmatic case of the captive audience from the passengers on the public buses or the child running through stations on the radio dial,(102) to the employee working for low wages in a tight job market who is sexually harassed by her employer or co-worker. This shift in paradigmatic understandings, inspired by the legal realist analysis of will and duress, might do much to set first amendment law on the right path.(103)

Finally, integrating issues of sexual and racial harassment into first amendment law will require us to rethink the doctrines of vagueness and overbreadth, which for so long have served libertarian interests. For example, the Equal Employment Opportunity Commission’s definition of sexual harassment includes “verbal … conduct of a sexual nature … when … (3) such conduct has the purpose or effect of unreasonably interfering with an individual’s work performance or creating an intimidating, hostile, or offensive working environment.”(104) If a statute involving loitering or breach of the peace used language of such generality, I suspect that most left-libertarians immediately would pronounce it overbroad and vague. My argument is not that regulations conforming to the EEOC’s guidelines are necessarily unconstitutional (or that no limiting constructions are available). Rather, I suggest that this is yet another example in which the tools of analysis that have served the left-libertarian position on free speech so well in the past are ill-adapted to the problems of the present era. Sexual harassment statutes are not the same as loitering statutes and breach of the peace statutes. The type of analysis required must differ because the subject matter differs.(105)

Offhand, I can think of two ways in which the analysis of overbreadth and vagueness in the context of a loitering statute might differ from that involved in regulations against racial or sexual harassment. First, the remedies offered for violation of the statue or regulation are quite different. We should be more concerned about imposing criminal sanctions on the unwary than ordering back pay and reinstatement against an employer in a close case. Second, and more importantly, the issues of power in the paradigmatic situations in which the two regulations are enforced are quite different. The danger in the loitering case is that a more powerful entity (the state) will take advantage of vagueness or overbreadth to punish persons who are unpopular or unorthodox in their appearance, manner, dress, or behavior. In the case of the harassment regulation, the danger is that a vague or overly broad statute will chill conduct by the more powerful party (the employer or the co-worker who harasses). To be sure, too great a chilling effect will be deleterious to employer-employee relations. Certainly it would be unfortunate if superiors were continually worrying whether the last thing they said to their subordinates will be misconstrued and precipitate a lawsuit. But these possibilities for abuse of power are less troublesome than an abuse of power by police officials who have a monopoly on the use of force against, for example, homeless citizens who have no effective recourse against arbitrary law enforcement.

One might object that whatever the force of these arguments, traditional first amendment doctrine at least has the advantage of content neutrality in applying the doctrines of overbreadth and vagueness. Yet I suggest that this form of neutrality, like so many others in first amendment law, was always illusory. For example, the Supreme Court has recognized that substantial overbreadth is required for facial invalidity.(106) Left-libertarians have understood this doctrinal move by the Burger Court as a serious threat to first amendment rights, for “substantiality” is a sufficiently loose concept that courts will be given considerable leeway in determining what constitutes substantial overbreadth. Yet even Justice Brennan’s dissent in Broadrick v. Oklahoma recognized that the Court “ha[s] never held that a statute should be held invalid on its face merely because it is possible to conceive of a single impermissible application, and in that sense a requirement of substantial overbreadth is already implicit in the doctrine.”(107) What separates Justice Brennan’s insight from Justice White’s majority opinion is the degree of substantiality required for invalidity. And this is not a question that can be determined simply by counting up possible hypothetical applications, even if that task were possible. The question of substantial overbreadth is one of quality as well as quantity. This places the courts in the unenviable position of making judgments of value and context, but this task is no less necessary if one subscribes to Justice Brennan’s position. In order to accept Justice Brennan’s arguments, one must agree that the examples he gives of protected conduct reached by the statute in question are substantial,(108) and that the type and degree of expression chilled by the statute are also substantial. There are ways of deciding these questions, but they are not indisputable or mechanical, and they are certainly not neutral or value-free.(109)

You will note that in my earlier analysis relations of economic and social power figured prominently. In assessing what constitutes substantial overbreadth or vagueness, I do not think it inappropriate to employ common sense judgments about the way the world works. Although the distinction between public power and private power is significant, even more significant for me are what power relations (public or private) exist in the standard case in which the statute operates. I do not claim that this sort of analysis is formally neutral or value-free. Indeed, eschewing claims to this sort of “neutrality” is the only way one can acknowledge that being a homeless indigent rounded up by police in routine sweeps and being an affluent, white, male employer accused of making passes at an underpaid and overworked female secretary are two very different sorts of experiences. I simply am making overt the sort of matters of judgment that courts must make in any case when determining if a statute is substantially overbroad or vague – the degree to which protected conduct will be chilled and the nature and significance of the conduct likely to be chilled. Certainly, the above judgments I have offered about power relations are fallible. One might debate them. But that, of course, is really the point – one should be permitted to debate these issues openly in order to decide what is or is not substantially overbroad and vague.(110)

Moreover, when I say that issues of vagueness and overbreadth are matters of degree and context, I am not making an argument against having any general rules of construction in first amendment cases. Rather, I suggest that we ask ourselves what types of paradigmatic situations call for more and less tolerance in assessing overbreadth and vagueness. Once we have identified contexts in which vagueness and overbreadth concerns are more important – say loitering statutes – and less important – say the workplace – then we can use rules to give judges some direction about how to apply these concepts. Indeed, we do this already. We are much more concerned with vagueness in criminal statutes affecting expressive activity than we are with vagueness in statutes that do not touch upon speech. This same form of reasoning should also hold true within the class of statutes touching upon expressive activity. The legal realists taught us to be suspicious of overly broad abstractions in our legal concepts and sensitive to alterations of social context. They did not, however, suggest we abandon the policy of using rules to give direction to decisionmakers and to ease administrative burdens.

I thus conclude that precedents like Broadrick, as well as the Burger and Rehnquist Courts’ increasing predilection to decide first amendment cases on an as-applied basis rather than through facial challenges,(111) may be of increasing importance to litigators who seek to protect harassment-in-the-workplace statutes from constitutional invalidity. It is undoubtedly strange and ironic that liberals who decried cases like Broadrick in the 1970s should now employ similar rationales in the 1990s. This, however, is simply another example of ideological drift. A doctrine of law takes its meaning from the contexts in which it is applied repeatedly. Thus the political meanings of the doctrines of overbreadth and vagueness shift in political valence as they are used repeatedly in new historical contexts.


This brings me, at last, back to the first amendment problem with which I began this Article – the constitutionality of Kansas City’s abolition of a public access channel to keep the Ku Klux Klan off the air. When I first wrote my memo to the City Council, I saw this problem solely as an issue of content neutrality – the City may abolish the public access channel for many reasons, but it may not do so to keep a particular speaker from speaking or a particular viewpoint from being heard. I now see the issues differently. For me, this case poses in striking fashion two conflicting interests for the left in contemporary first amendment law. The first is the need to ensure effective access to the means of communication for all points of view, including unpopular ones; the second is the important state interest in eliminating racial discrimination and protecting racial minorities from harassment and abuse.

My present view is that the City’s action is still probably unconstitutional. My reasons for thinking so, however, are somewhat different than before. First, if the City signs a monopoly agreement with a particular cable television company, I think that the City probably is obligated to ensure that there is a public access channel available to all on a first come, first served basis.(112) Even if I am wrong in this conclusion, I believe that once having provided such a public forum, the City may not withdraw it unless it can provide very strong justifications; withdrawal because of distaste for the messages conveyed on the channel is not a sufficiently good reason.

One such sufficiently good reason for limiting access might be harassing behavior against a captive audience. If the City’s grant of access assisted the Klan in assaultive behavior, akin to direct face-to-face racial harassment, there might be a justification for a limitation on access used for this purpose. On the other hand, there is no reason to think that the showing of the series “Race and Reason” was equivalent to such a face-to-face verbal harassment. Moreover, being the sort of person who is not inclined to assume that movies I have not seen or books I have not read contain materials in need of censorship, I would not be willing to restrain a showing of “Klansas City Live” before the fact, unless it were proved that it would be used as a forum for assaultive and harassing behavior. I think this would be very difficult to prove. It may be that the medium of television, by its very nature, is rarely as assaultive as a face-to-face encounter, but I do not think it necessary to decide that issue as a matter of law. Even if it could be proved that the Klan’s behavior on live television rose to that level, I think the appropriate remedy would be to restrain them from that type of behavior alone and not from other racist but non-assaultive or harassing speech they might happen to offer on their program.(113)

This is an admittedly preliminary attempt at resolving the difficult first amendment issues presented in the Klan Cablevision case based upon the framework outlined in this Article. Perhaps others using a similar analysis might arrive at different conclusions. I would be surprised indeed if the suggestions I have offered led to a simplistic, mechanical jurisprudence, or avoided the need for difficult moral and political choices. Rather, I offer my analysis because I believe that the problems of the future cannot be solved using the intellectual frameworks of the past, no matter how much good they may have done us. Progress in politics and in law is not simply a matter of convincing others to think as you do. It also requires having the courage to change your own ways of thinking when changing times require it.


68. My pun on the familiar Marxist term “means of production” is both deliberate and ironic. It is deliberate because communication is like production in that it requires investment in certain forms of property. It is ironic because belief in the power of the means of communication ultimately rests upon a rejection of a materialist conception of history. Control of the means of communication matters because ideas matter, and ideas matter because they influence people to do things they would otherwise not do, regardless of the present ownership of the means of production.

69. Roberts has justly been accused of conveniently overlooking the previous history of regulation of public protest. See, e.g., Kairys, supra note 22, at 144. For an example of the earlier treatment of so-called “traditional” public forums, see Massachusetts v. Davis, 162 Mass. 510, 511, 39 N.E. 113, 113 (1895) (Holmes, J.), aff’d, 167 U.S. 43 (1897) (“For the Legislature absolutely or conditionally to forbid public speaking in a highway or public park is no more an infringement of the rights of a member of the public than for the owner of a private house to forbid it in his house.”).

70. For an important recent analysis of communicative technology, see I. DE SOLA POOL, TECHNOLOGIES OF FREEDOM (1983) (arguing against government regulation of the broadcast media).

71. What we think of as individual effort “unaided” by economic power is often most effective at the beginning of a new technology. For example, everyone is familiar with stories of the computer hacker whose software program catapulted him or her to millionaire status in a matter of months. But as the computer industry matures, these sorts of success stories occur less and less frequently. The latest versions of the standard commercial software products now require armies of programmers, substantial expenditures on advertising promotions, and all of the other requirements of modern corporate marketing and production. Moreover, it may even make economic sense for existing market participants to engage in behavior that increases barriers to entry by new competitors, at least after an industry has reached a given level of maturity.

72. Fiss, Why the State?, 100 HARV. L. REV. 781 (1987); Fiss, Free Speech and Social Structure, 71 IOWA L. REV. 1405 (1986). Both of these essays are highly recommended for their fresh approach to the problems of first amendment law. For a spirited rejoinder, see Powe, Scholarship and Markets, 56 GEO. WASH. L. REV. 172 (1987).

73. Note that this argument about scarcity should be distinguished from the claim of “scarcity” made in Red Lion Broadcasting Co. v. FCC, 395 U.S. 367 (1969) – that regulation of broadcasting could be justified by natural limits to the number of persons who could broadcast simultaneously over the airwaves. As Professor Powe notes, the justification of content based regulation on the basis of a factual claim of comparative scarcity is spurious, given the much larger number of television and radio stations than newspapers. Powe, “Or of the [Broadcast] Press,” 55 TEX. L. REV. 39, 55-56 (1976). One might also note that despite fears of comparative scarcity, many cable, VHF, and UHF channels remain unused or underutilized in most communities. See I. DE SOLA POOL, supra note 70, at 138-54. My point is that all forms of communication are scarce to the extent they involve the expenditure of resources and control of communicative technology. It is the general scarcity of things of value, and not the particular scarcity of the medium of broadcasting, which is my concern here.

74. See Coase, The Federal Communications Commission, 2 J.L. & ECON. 1, 36 (1959); Fowler & Brenner, A Marketplace Approach to Broadcast Regulation, 60 TEX. L. REV. 207, 256-57 (1982).

75. A different form of scarcity of access always existed because of the limited number of persons one could communicate with in a particular amount of time.

76. See Hohfeld, supra note 16, at 32-33.

77. 347 U.S. 483 (1954); see Wechsler, Toward Neutral Principles of Constitutional Law, 73 HARV. L. REV. 1, 34 (1959).

78. Hale, Coercion, supra note 13, at 470-76. Peller’s treatment of Hale’s arguments, Peller, supra note 13, at 1232-40, is especially good. See also Samuels, The Economy as a System of Power and Its Legal Bases: The Legal Economics of Robert Lee Hale, 27 U. MIAMI L. REV. 261 (1973) (summary and exposition of Hale’s work).

79. Thus, even state laws regarding theft and destruction of property affect access to the means of communication. One can communicate by commandeering a television station and holding the station managers hostage until they deliver one’s message. Indeed, one can also communicate a message by killing someone, or by blowing up a building. That is one reason why acts of political terrorism are performed. (Another is that antisocial behavior gains media attention – it is the poor person’s way to gain access to the mass media.) Of course, we do not allow people to engage in acts of political terrorism, and we are quite right to forbid them, even if such prohibitions have an incidental impact on the communicative power of private parties. We also do not allow newspapers to cut down trees to make paper unless they purchase the appropriate property rights. Nor, I suspect, would we allow even President Bush to seize a flag factory during a Presidential campaign in order to convey a patriotic message. My point is simply that we need not look very far for restraints on our communicative powers. They are all around us, in the social and economic structures the law sustains and enforces. The state is always granting and denying access to the means of communication through its distribution of economic power to private parties, power that in turn can be used by private parties to grant or deny such access.

80. More commonly, marketplace forces require the mass media to cater to the great mass of public tastes. Thus the mass media tend to reinforce mainstream ideas – the path of least audience resistance. To this end, they generally offer radical ideas on the left and right only as the intellectual equivalent of a freak show, thereby strengthening our faith in mainstream thought by displaying unusual ideas and opinions as things to ogle and marvel at.

81. FCC v. Pacifica Found., 438 U.S. 726 (1978) (FCC could prohibit certain types of offensive speech on the airwaves because persons receiving such broadcasts in their homes are in position of captive audience); Lehman v. City of Shaker Heights, 418 U.S. 298, 305 (1974) (Douglas, J., concurring) (city may ban political advertising on its buses because commuters are captive audience); Rowan v. Post Office Dep’t, 397 U.S. 728 (1970) (first amendment does not protect right to send unwanted material into home of another).

82. A point that Justice Harlan explicitly recognized in his opinion in Cohen v. California, 403 U.S. 15, 21 (1971): “The ability of government … to shut off discourse solely to protect others from hearing it is … dependent upon showing that substantial privacy interests are being invaded in an essentially intolerable manner.” The very words used – “substantial” and “essentially intolerable” – indicate that this question does not admit of precise answers.

83. Hale, Bargaining, supra note 13, at 625-28; Hale, Coercion, supra note 13, at 473-77; Samuels, supra note 78, at 302-09.

84. One might try to avoid these conclusions by defining coercion as an attempt to force a person to do what she is not legally required to do by means one is not legally entitled to use. In that case, the problem of coercion disappears because one never coerces by definition when one is acting within one’s rights and never fails to coerce when one is acting outside of them. However, Hale pointed out, this approach creates a problem of circularity. For what concerns us in assessing the justness of the law is whether rules of law unfairly allow parties to coerce each other. But the above definition of coercion would indicate that no matter what system of rules we had, the law never sanctions coercion. See Hale, Coercion, supra note 13, at 476; see also Peller, supra note 13, at 1235-36. Thus, Hale’s point is that a purely positivistic definition of coercion (defined in terms of legal rights) is either empty or circular.

85. See Peller, supra note 13, at 1237-39.

86. Note the difference between the circumscription produced by physical conditions and that produced by legal rules. For example, if a person desires to be the world’s greatest sprinter but has only one leg, we do not say that she lacks free will or freedom to choose. It is nevertheless true that what we call “exercises of her will” are circumscribed by her physical condition. The difference between this case and the critique of formal freedom of contract is that we are concerned with defending legal rules as just or unjust, but not physical conditions. The argument about free will exercised in the context of physical limitations would be circular only if it were within our power to alter those conditions.


Thus, if technology existed miraculously to endow the would-be sprinter with a perfectly functioning leg and the talents necessary to become a world class athlete, ownership of that technology would be determined by the existing legal rules of property and contract. At that point an issue of justice would arise as to whether the sprinter should be entitled to that technology, and under what conditions. We can see this better by choosing a less fanciful example. Suppose a cure for a previously incurable disease (say AIDS) has been discovered, but the right to distribute the drug is held privately. If the regime of contract and property rules results in some AIDS victims not purchasing the drug because they cannot not afford it, it would be circular to argue that this result was morally just merely because the failure to contract was a result of those AIDS victims’ free will – that they freely chose from among the best of the legally available alternatives. This is not to say, however, that the justness of a particular distribution of medical technologies might not be established on grounds other than the concept of free will exercised in the marketplace.

87. See, e.g., Erznoznick v. City of Jacksonville, 422 U.S. 205 (1975) (limited privacy interest of persons on the streets places burden on them to avoid offensive expression); Cohen v. California, 403 U.S. 15, 21 (1971) (“Those in the Los Angeles Courthouse [who objected to message on defendant’s jacket] could effectively avoid further bombardment of their sensibilities simply by averting their eyes.”).

88. Although it would take me too far afield to discuss the matter at length, I should point out that the problem of coercion in education is part of a larger issue – namely the use of the right of free speech as a means of ideological control. In my use of the analytical framework developed by the legal realists I have assumed that one is partly free and partly coerced when one chooses what one thinks best given the limitations created by legal rules. But this argument assumes that one also determines “what one thinks best” – that is, that one determines one’s own values. Yet control of the means of communication is also an important method of shaping and altering the values of listeners. This is especially true in mass communication, where the audience listens but has no opportunity to talk back. The unanswered messages conveyed may have the effect of normalizing and naturalizing particular attitudes and beliefs. M. POSTER, FOUCAULT, MARXISM AND HISTORY: MODE OF PRODUCTION VERSUS MODE OF INFORMATION 115 (1984).


A normalizing process, of course, always has existed in education of the young. We want our children to accept the values we teach them in schools. Indeed, the right to instill values may be even more effective than rights to coercion through the use of legal rules, since values internalize restraints upon behavior. Thus, although we believe that “brainwashing” is bad and inimical to notions of individual self determination, we nevertheless simultaneously believe that some forms of ideological control may be justified “for our own good,” or for the good of society. The difficulty is heightened when there is the danger that ideological control may be used to perpetuate relations of power that are thought undesirable. See generally C. MACKINNON, supra note 7 (perpetuation of private power); M. YUDOF, supra note 20 (public power).


Thus, the right of free speech soon runs headlong into another right that the left seeks to foster – the right to education. The right to education is the right to particular forms of training and cultural indoctrination, but such training and such cultural indoctrination may raise difficult first amendment issues of ideological coercion and control. See, e.g., Board of Educ. v. Pico, 457 U.S. 853 (1982) (first amendment restricted ability of school library to remove books thought offensive); Bethel School Dist. No. 403 v. Fraser, 478 U.S. 675 (1986) (school had right to regulate offensive speech given at student assembly); Hazelwood School Dist. v. Kuhlmeier, 484 U.S. 260 (1988) (school principal had right to edit student newspaper as part of school’s educational mission of instilling respect for appropriate values).

89. 198 U.S. 45 (1905).

90. 236 U.S. 1 (1915).

91. See Erznoznick, 422 U.S. at 206-07.

92. See Lawrence, supra note 4.

93. 315 U.S. 568 (1942).

94. 395 U.S. 444, 447 (1969).

95. See supra note 16 and accompanying text.

96. See Strauss, Sexist Speech in the Workplace, 25 HARV. C.R.-C.L. L. REV. 1, 3 n.12 (1990).

97. 42 U.S.C. §§ 2000e to 2000e-17 (1988).

98. See Meritor Savings Bank, FSB v. Vinson, 477 U.S. 57, 77 (1986) (Marshall, J., concurring); Note, Relief for Hostile Work Environment Discrimination: Restoring Title VII’s Remedial Powers, 99 YALE L.J. 1611, 1613-19 (1990). Although one could obtain legal remedies under 42 U.S.C. § 1981 (1988), it does not reach sex discrimination, and the Supreme Court has now decided that it does not encompass claims of racial harassment. Patterson v. McLean Credit Union, 109 S. Ct. 2363 (1989). The proposed Civil Rights Act of 1990, H.R. 4000, 101st Cong., 2d Sess. (1990); S. 2104, 101st Cong., 2d Sess. (1990), vetoed by President Bush, see 136 CONG. REC. S16562 (daily ed. Oct. 24, 2990), would have reversed the Supreme Court’s holding in Patterson by allowing causes of action for racial harassment under § 1981. The Act would also have authorized traditional legal remedies, including compensatory and punitive damages, for violations of Title VII.

99. See Delgado, supra note 6. For a discussion of the interaction of the tort of intentional infliction of emotional distress with the first amendment, see Anderson, Tortious Speech, 47 WM. & MARY L. REV. 71 (1990); Austin, Employer Abuse, Worker Resistance, and the Tort of Intentional Infliction of Emotional Distress, 41 STAN. L. REV. 1 (1988); LeBel, Emotional Distress, the First Amendment, and “This Kind of Speech”: A Heretical Perspective on Hustler Magazine v. Falwell, 60 U. COLO. L. REV. 315 (1989); Love, Discriminatory Speech and the Tort of Intentional Infliction of Emotional Distress, 47 WM. & MARY L. REV. 123 (1990); Post, The Constitutional Concept of Public Discourse: Outrageous Opinion, Democratic Deliberation, and Hustler Magazine v. Falwell, 103 HARV. L. REV. 603 (1990); Smolla, Emotional Distress and the First Amendment: An Analysis of Hustler Magazine v. Falwell, 20 ARIZ. ST. L.J. 423 (1988); Smolla, Rethinking First Amendment Assumptions about Racist and Sexist Speech, 47 WM. & MARY L. REV. 171 (1990); Woman, Verbal Sexual Harrassment on the Job as Intentional Infliction of Emotional Distress, 17 CAP. U.L. REV. 245 (1988).

100. See Note, A Communitarian Defense of Group Libel Laws, 101 HARV. L. REV. 682 (1988).

101. For a thoughtful attempt, see Strauss, supra note 96.

102. See FCC v. Pacifica Found., 438 U.S. 726 (1978); Lehman v. City of Shaker Heights, 418 U.S. 298 (1974).

103. I have so far emphasized the problem of coercion in speech situations, and one of my previous examples involved racist speech on a college campus. The subject deserves a much fuller treatment than I can offer here; however, I should note that the problems of racist speech in university settings involve somewhat different considerations than the workplace. First, the university setting raises distinct issues of privacy as well as coercion, especially because students live on college campuses as well as work there. Second, educational systems do have a necessarily inculcative purpose, which means that universities may have a legitimate interest in fostering certain types of values – for example, they may have a legitimate interest in instilling values of respect and tolerance for different persons and for ideas different from one’s own.


These additional interests do not justify blanket prohibitions on racist and sexist speech. For example, the inculcative interest cuts both ways – it may require some deterrence of intolerant expression, or it may require enforced toleration of the intolerant. Cf. L. BOLLINGER, THE TOLERANT SOCIETY 237-48 (1986) (first amendment protection of unpopular speech necessary to instill virtues of tolerance in society as a whole). Rather, the students’ interest in privacy and freedom from coercion, and the university’s inculcative interests counsel that universities must be all the more sensitive to the specific contexts in which speech occurs, and to the competing interests involved. What would be too great an invasion of privacy in the dormitory (a racist poster slipped under a student’s door, for example), may have to be treated differently from offhand remarks in the cafeteria, comments made in the classroom, or arguments in the public streets outside the campus. In addition, the coercive nature of speech may differ in each of these places.


Note that if coerciveness were our only concern, racist and sexist statements would gain no additional protection from being expressed in the classroom. Because students must attend classes, the classroom can easily present as coercive a situation as the workplace, even if the students’ privacy interests are minimal in comparison to the college dormitory. However, the university’s legitimate inculcative interest in tolerance and respect for dissenting views is important in the classroom in a way that it is not in the workplace. This suggests that the classroom and the workplace cannot be treated alike in all respects, nor can the classroom and the dormitory.

104. 29 C.F.R. § 1604.11(a) (1989). The full text of the EEOC guidelines is as follows:


(a) Harassment on the basis of sex is a violation of Sec. 703 of Title VII. Unwelcome sexual advances, requests for sexual favors, and other verbal or physical conduct of a sexual nature constitute sexual harassment when (1) submission to such conduct is made either explicitly or implicitly a term or condition of an individual’s employment, (2) submission to or rejection of such conduct by an individual is used as the basis for employment decisions affecting such individual, or (3) such conduct has the purpose or effect of unreasonably interfering with an individual’s work performance or creating an intimidating, hostile, or offensive working environment.

Id. (footnote omitted).

105. For an excellent discussion of the context-sensitive nature of overbreadth analysis, see Redish, The Warren Court, the Burger Court, and the First Amendment Overbreadth Doctrine, 78 NW. U.L. REV. 1031 (1983).

106. See Broadrick v. Oklahoma, 413 U.S. 601, 615 (1973).

107. Id. at 630 (Brennan, J., dissenting).

108. See id. at 628.

109. What I have said about overbreadth can be applied, in similar fashion, to the problem of vagueness. If there is no specific doctrine of “substantial vagueness” in first amendment law, it is because the requirement of substantiality has always been understood. Whether words are sufficiently unclear that persons of common understanding must guess at their meaning is a matter of degree, of practical judgment. All language is clear and all language is vague, depending upon the circumstances and the degree of precision required by those circumstances. Courts have no more value-free methods of determining vagueness than they do of determining overbreadth.

110. Cf. Redish, supra note 105, at 1069-70 (properly performed, overbreadth analysis requires sensitivity to context that cannot be provided by broad categorical rules).

111. E.g., Massachusetts v. Oakes, 109 S.Ct. 2633, 2638 (1989); Brockett v. Spokane Arcades, 472 U.S. 491, 504 (1984); United States v. Grace, 461 U.S. 171, 175 (1983).

112. This is now required by statute in some cases. See 47 U.S.C. § 532(b)(1) (1988) (commercial (non-governmental) access must be granted for a given number of cable channels depending upon number of total channels available as specified in franchise agreement; franchise owner may exert no editorial control over content of programming on such channels). I would argue a grant of access is also a constitutional requirement. For an argument that monopoly grants to cable franchises are themselves unconstitutional, see L. POWE, AMERICAN BROADCASTING AND THE FIRST AMENDMENT 239-47 (1987). The statutory requirement of provision of a public access channel apparently did not apply in the Kansas City case, either because the number of channels was too small or because the original franchise contract was entered into before the statutory requirements took effect.

113. Professor Matsuda has argued for the right to regulate non-assaultive, non face-to-face and overtly political racist speech. Language as Violence, supra note 4, at 361-64 (remarks of Professor Matsuda); Matsuda, supra note 6. From what I have said, it should be clear that I do not believe that regulation of racist speech can or should go so far. My analysis suggests that forms of coercion may nevertheless result from allowing such speech in our communities. But I would require somewhat more proof before I was convinced that the degree of coercion in racist political speech is sufficiently different in degree and kind from that produced by other forms of political speech – for example, the coercion produced by an anti-abortion protester who shouts pro-life slogans outside an abortion clinic.