Universities committed to the value of learning must be committed, above all, to the free exchange of ideas. But it does not follow that they are obliged to accept any and all forms of speech. Under a system of free expression, universities -- and, indeed, the government -- can legitimately distinguish between verbal assaults and speech that forms part of the exchange of ideas. Those who deny this distinction will, I think, have trouble making sense of their own deepest beliefs about the kinds of speech that enjoy full protection of the First Amendment and those that do not.
In opposing restrictions on speech, First Amendment absolutists cite the exact Ianguage of the First Amendment: "Congress shall make no law abridging the freedom of speech." But the bare words will not help us to decide what amounts to an "abridgement" and what falls within the category of "freedom of speech." The history of the amendment shows that the text was not thought to protect all speech under all circumstances. Indeed, no one has ever understood it in this way.
Even today, when the First Amendment is understood to provide broader protection than ever before, restrictions on speech abound. The government regulates perjury; bribes; threats; unlicensed legal, medical, or psychiatric advice; fraud; libel of private citizens; libel of public officials; obscene or harassing telephone calls; child pornography; false or misleading commercial speech; conspiracies; obscenity; the dissemination of technological information with potential military applications; the disclosure of names of agents of the Central Intelligence Agency (CIA); criminal solicitation; sexual harassment. Even those who disagree with rules restricting one or more of these forms of speech will find it hard to argue that we should not regulate any of them.
But how are we to decide which sorts of speech fall outside the Constitution's protection? What principles underlie the distinction between constitutionally protected speech and speech that can be proscribed?
Some tempting solutions come to mind. Perhaps the state can regulate speech only when, on closer inspection, speech turns out to be action. But child pornography, psychiatric advice, misleading commercial speech, and conspiracies are speech, not action. If they can be regulated, the reason lies in their distinctive properties and consequences. The word "action" is just a placeholder for an analysis of those distinctive features.
Perhaps the state can regulate speech when, and only when, it will produce certain and imminent harm. But this solution also fails, at least as a complete answer. Attempted threats and bribes, criminal solicitation, and attempts to conspire are often pitifully ineffective and therefore harmless, but still properly subject to penalty of law. We restrict deceptive commercial speech, but not deceptive political speech; but it is hardly clear that deception in commerce is always more harmful than deception in politics. Indeed, the opposite may be true. In evaluating whether speech can be restricted, something other than harm must be at work.
That something, I suggest, has to do with the relationship between speech and the values for which the principle of free speech should be taken to stand. The "core" value, I propose, is rational deliberation, especially about democratic matters. If speech contributes to the exchange of ideas about some question of public importance or is part of rational discussion about a problem, it ought to fall unambiguously within the "core" of free speech protected by the First Amendment. As a rough start, this principle helps to explain why the kinds of speech now deprived of constitutional protection are placed in a different category. Bribery, fraud, libel, perjury, pornography, sexual harassment, and so on are deemed outside the sphere of protected speech not merely because they are harmful, but also because they have an attenuated connection to the values that underlie the free speech guarantee.
To ensure large scope for democratic discussion, we must interpret the basic standard broadly. We should insist that some forms of speech fall within the "core" regardless of their emotive qualities, which are, in fact, inseparable from their substantive content. And whether or not speech falls within the core, government must always have a legitimate justification for regulation, that is, a justification that points to something other than government's skepticism about the viewpoint of the speaker. Commercial speech can be regulated if it is false and misleading, and thus threatens genuine governmental interests; legal advice can be banned if the speaker is not a member of the bar; but in no case can government regulate speech simply because it disagrees with what someone has to say. In most cases in which we are genuinely concerned about censorship, this kind of motivation turns out to be at work.
Regulation of Hate Speech
To make sense of the current controversy over "hate speech," we need to make some distinctions. Hate speech ranges from face-to-face verbal assault, to threatening mail, to broad statements about race, gender, and sexual orientation that seem to many offensive, hurtful, and hateful.
If, as I argue, government should not regulate speech that contributes to the exchange of ideas, it should not regulate speech that is, or seems, sexist or racist when that speech is a part of the exchange of ideas. To allow government to censor racist or sexist statements because they are offensive or even harmful would be wholly inconsistent with our system of freedom of expression. This is sufficient reason to disapprove of the University of Michigan's judicially invalidated ban on "[a]ny behavior, verbal or physical, that stigmatizes or victimizes an individual on the basis of race, ethnicity, religion, sex, sexual orientation, creed, national origin, ancestry, age, marital status, handicap, or Vietnam-era veteran status, and that . . . creates an intimidating, hostile, or demeaning environment for educational pursuits...." This sort of broad ban includes a wide range of statements that are part of the exchange of ideas, and it fails to give people sufficient notice of what statements are allowed.
On the other hand, racial, religious, or sexual epithets are sometimes used in a way that makes it exceedingly hard to consider them a contribution to the exchange of ideas. Rather, they are put forth as an assault or a threat, corresponding to an obscene telephone call. Although some lawyers may find that distinction, in principle, to be elusive or incoherent, those interested in working through hard questions will find it not only usable, but indispensable for thinking through the problems that hate speech raises.
We might compare the Michigan approach with Stanford's narrowly drawn rule against "harassment by personal vilification." To fall within that category, three independent findings are required. The speech must be (1) "intended to insult or stigmatize an individual or a small number of individuals on the basis of their sex, race, color, handicap, religious, sexual orientation, or national and ethnic origin"; (2) addressed "directly to the individual or individuals whom it insults or stigmatizes"; and (3) "make use of 'fighting words' or non-verbal symbols," defined as words or pictures that "are commonly understood to convey direct and visceral hatred or contempt for human beings" on one of the enumerated bases. This standard, based on Supreme Court decisions that free expression does not cover all use of "fighting words," is rooted in an objection not to a point of view, but to assaultive or threatening behavior.
The Stanford rule might be thought un-acceptably selective, since it does not regulate epithets ("fascist," "pig," "communist," "moron") other than those involving certain enumerated grounds. But a university could reasonably conclude that the subjective harm of face-to-face insults based on race or sex is usually more severe than that of insults founded on some other ground; that a more general prohibition on epithets would threaten to spill over into valuable or protected speech (consider "fascist" or "communist"); and that the grounds specified by Stanford have sufficient clarity and definition, as other grounds do not, to stand as a basis for regulation. The fact that what counts as "insult" or "stigmatization" will depend on both time and place is not a good objection to this approach. An intent to insult or stigmatize is the necessary criterion, and any reasonable administration should apply its restrictions with sensitivity and restraint. For these reasons, the Stanford solution should stand as a model for what a reasonable university might do, just as the Michigan approach should stand as a warning about the possibility of excess.
Suppose a state legislature enacted a prohibition like Stanford's. Here are the questions I have for those who would oppose that prohibition: On what rationale is it possible simultaneously to approve of the many restrictions on speech that are so common in our law? Must a democratic society tolerate obscene telephone calls? If not, what is the difference between that form of harassment and the forms that our hypothetical state legislature has undertaken to regulate?
The Campus, Public and Private
The university campus has several relevant differences from government. First, universities are uniquely dedicated to the exchange of ideas; that is what they are for. Second, the university is necessarily a restricted setting; it is a community, whose members are, in principle, free and equal. The university has a special duty to combat the harmful effects of certain conduct and words; it has a legitimate interest, consistent with free expression, in protecting its members from the fear of violence, feelings of second-class citizenship, and even coerced silence. Words that threaten physical intimidation can indeed produce silence. Much of First Amendment absolutism in this context reflects an absence of empathetic identification with the victims of hatred or contempt on campus and elsewhere.
Finally, and perhaps most important, most universities are private rather than public, and not covered by the Constitution at all. The framers' deliberate decision to exempt the private sector from constitutional limits serves liberty well; it allows for a desirable degree of diversity, flexibility, and experimentation. Different norms and practices make perfect sense when we are thinking about diverse institutions.
Taken as a whole, these considerations suggest, first, that a university, public or private, should not adopt vague and broadly worded prohibitions, such as those adopted by the University of Michigan. That does not mean, however, that the federal government should force all universities into any particular mold. The recent proposal of the American Civil Liberties Union (ACLU) and Congressman Henry Hyde, Republican of Illinois, to apply federal constitutional limits to private universities would be a grave mistake, one that is ironically inconsistent with the commitment of Congressman Hyde and the ACLU to freedom and diversity. In a pluralistic society, it is probably desirable that some universities are experimenting with more intrusive forms of regulation even if those forms of regulation are unjustified. It will be useful for all of us to see the consequences of such experimentsto see what kinds of effects they have on the exchange of ideas and the lives of the students and teachers involved. But to oppose a federal ban on experiments is not to endorse those experiments.
There should be no objection, by contrast, to a measure like Stanford's rule against "harassment by personal vilification." If such a measure would be acceptable when enacted by the state, it is surely acceptable in the setting of the university. Indeed, in the university setting the Stanford restriction may be somewhat too narrow. Universities necessarily regulate speech. A professor who teaches biology can be punished for instructing students in mysticism. A student who insists on using vulgar language in class can be asked to leave. A professor or student who uses the classroom as an occasion for hurling insults can be disciplined. In this respect, restrictions on academic expression are unavoidable and appropriate.
The inevitability of some sort of restriction -- together with the special responsibility of the university for its members -- suggests that it may be desirable for some universities to reach beyond the Stanford solution. Much will depend on the particular circumstances. A university that has experienced a number of incidents of hatred -- threatening mail to gay students, the posting of violent pornography on the doors of women students, epithets directed at blacks -- must take the hate speech issue more seriously than one that has not. The norms of the particular place matter as well.
Although I personally favor the Stanford approach for most institutions, it is at least reasonable to think that some universities should also ban speech that is not directed at particular individuals, but instead at certain groups as a whole when it reflects pure visceral contempt and is not plausibly intended or received as a contribution to discussion. Such limitations should be defined narrowly, and none should approach the breadth of Michigan's effort. But it seems plausible to allow some experimentation in this direction. Some universities, for instance, should take special steps to protect women students from threatening speech that associates sex with violence. In every case, any decision to punish students should be undertaken with a firm guarantee of procedural fairness, a sensitive understanding of the context, and an awareness that informal and consultative measures are usually far preferable to suspension and punishment.
Today the regulation of hate speech raises an objection I have not yet considered: that this regulation is really part of an effort to impose a political orthodoxy, and to use any means -- persuasion, social stigma, custom, and the force of law -- to do so. I agree that we should resist any effort to impose an orthodoxy through "codes" intruding on the exchange of ideas. The rules I have defended impose no orthodoxy; they are designed only to serve as an acceptably neutral means of protecting students from behavior that violates the premises of education and the general standards governing how students, faculty, and administrators should relate to one another. Supporters of such regulations need hardly ally themselves with opponents of the free exchange of ideas; on the contrary, that is the principle that lies behind their support of restrictions on verbal assaults.
I do not defend the "speech codes" of every college and university. Many of these are inconsistent with the commitment to the unfettered exchange of ideas, central both to the First Amendment and to academic life. But some rules promote that commitment. Within the boundaries I have set out, the guiding principle should be one of pluralism. No single rule will fit the diverse norms and practices of the thousands of public and private universities in the United States. So long as they respect the exchange of ideas, and so long as any restrictions are both narrow and clear, those universities should be free to experiment with different methods for controlling forms of abuse that have become an increasing problem for university life.