Book Review

By Michael E. Rosman

Constitutional Commentary, December 20, 1996

In Fighting Words: Individuals, Communities, and Liberties Of Speech, Kent Greenawalt surveys a number of different free speech issues all of which revolve around a common question: What happens when a person’s expression, through words or deeds, significantly disturbs another person or persons? The traditional civil libertarian answer to this question (with few exceptions) has been simple: Tough luck to those who are disturbed. But the traditional civil libertarian response has come under increasing attack in recent years both by those who perceive free speech as a means by which the dominant classes in society can continue to subjugate the rest, forestalling true equality, and by those who believe that dissident or iconoclastic expression should not undermine our traditional societal values.

It is to Mr. Greenawalt’s credit that he sees the connection between hate speech and flag burning, workplace harassment and nude dancing, obscenity and campus speech codes. n3 He cov- [*318] ers all of these areas well; although lawyers have a tendency to overestimate the accessibility of writing on legal issues, I think the discussions can be followed by the average layperson, but will not bore the First Amendment specialist. (I fall somewhere in between.) Mr. Greenawalt also provides a comparison to Canadian rulings on similar First Amendment issues, n4 and tries to develop a theme (which I discuss briefly in Part III of this review) in the beginning and at the end of the book that emphasizes consideration of “communities” in First Amendment jurisprudence.

Mr. Greenawalt is a professor of law at Columbia Law School. Even if the jacket (which provides that information) of my copy of the book had somehow been misplaced, it would not have been a difficult thing to guess. His analysis is thorough, well-reasoned, and moderate. He sees many sides to the questions. He is concerned that the Supreme Court did not pay sufficient attention to interests being served by the Congressional flag burning statute, struck down in United States v. Eichman, n5 interests that differed from those served by the Texas flag-burning statute struck down in Texas v. Johnson. n6 (He thinks Eichman was correctly decided, mind you. It’s just that the Court could have been more sensitive in its analysis.) He gives full consideration to each varying effort to defend hate speech laws, even some that probably do not merit such attention. His analysis of workplace harassment is so evenhanded, and so thorough in discussing the pro’s and con’s of each separate argument, that I am still not sure what his position is even after reading it several times. Finally, as I discuss shortly, his analysis frequently reflects the values of the class of which Columbia Law School professors are a part. [*319]

In Part I of this review, I consider Professor Greenawalt’s analysis of hate speech and workplace harassment rules. In Part II, I look specifically at Professor Greenawalt’s analysis of campus speech codes, and more generally at his efforts to consider more closely the role that government is playing in each different free speech context, an idea that I liked and wished he had developed more fully. Finally, in Part III, I briefly examine his analysis of the “individuals” and “communities” that are included in the subtitle of his book and conclude – as I think Professor Greenawalt does himself – that they are, at best, of only modest use in considering free speech issues.


In discussing what is loosely characterized as “hate speech” – the pernicious and derogatory use of bigoted epithets or symbols – Professor Greenawalt begins, as he should, with “fighting words.” n7 Greenawalt wants to expand the definition of “fighting words” so as to include words which are designed to humiliate and which would be an invitation to fight to the “average” person. Thus, even if racial taunts are slung towards small African-American children by large white adults, Greenawalt considers them “fighting words” through an “equalization of victim” principle. n8 Greenawalt’s admitted purpose is to separate the “fighting words” doctrine from its roots as a First Amendment exception designed only to avoid actual violence, and to broaden it to include words that are intended to, and which generally do, inflict deep psychological hurt.

The “equalization of victim” principle sounds interesting, although, as Greenawalt concedes, it is inconsistent with the Supreme Court decision in Gooding v. Wilson, n9 and stretches the rationale for prohibiting this type of speech to the point where [*320] placing it under the “fighting words” rubric is a bit misleading. Greenawalt also recognizes two more serious problems. First, the use of “hate speech” carries some expressive value with it, even if the “messages” are ones of racial or ethnic inferiority that most of us would find distasteful. Second, an “intent to humiliate” will frequently be difficult to discern, particularly if it must be the sole motivation for the speech. n10 Greenawalt suggests that we limit a civil prohibition to “targeted vilification,” where the speaker has sought out the victim and initiated contact for the purpose of humiliation and insult, much like laws prohibiting harassing telephone calls. n11

Of course, there is the little matter of R.A.V. v. City of St. Paul, n12 in which the Supreme Court unanimously held that a similar law did not pass Constitutional muster. A five-person majority held that, even assuming that the municipal ordinance in that case precluding the use of “hate speech” involved only “fighting words,” it created improper distinctions between “fighting words” which discriminated on the basis of viewpoint (permitting minorities to yell “fascist” and “Nazi” at their enemies, but not allowing those yelled at to respond in kind) and content (not reaching other fighting words because they are not insults based on race, ethnicity, or another forbidden characteristic). n13 A four-person minority found the statute unconstitutional simply because it reached beyond the “fighting words” category to protected speech. As the foregoing discussion should make plain, however, Greenawalt simply disagrees with both groups of justices; with the minority because they do not have a sufficiently expansive notion of “fighting words,” and with the majority for using “strict scrutiny” to assess a statute which reached words with little expressive value. The intent to vilify or “humiliate” a [*321] private citizen, according to Greenawalt, is sufficient to take speech outside of the carapace of the First Amendment. n14

There is, of course, a rather obvious irony in a law professor suggesting that words spoken with an intent to humiliate are entitled to little or no Constitutional protection. Perhaps we should consider laws which focus on excesses in the use of Socratic dialogue and “targeted humiliation” within law schools. No doubt Greenawalt would state that such speech is different; it has a didactic purpose and, consequently, a public importance. I would agree. But that only reflects who Greenawalt and I (and, presumably, most of the readers of this review) are: people who would never use the kind of epithets that were the subject of the statute in St. Paul and who would be the target of a Greenawalt hypothetical statute, people who simply do not believe in the underlying messages that such epithets convey. On the other hand, people who would use such epithets – to, for example, “browbeat” minorities back towards their “proper” place of subservience in our country – might wonder why it is necessary to publicly humiliate law students in front of their classmates.

In short, our understanding of what is “valuable” speech (even with an intent to humiliate) tends to reflect our own backgrounds and values. And that is precisely the opposite of what a value-neutral First Amendment requires, and precisely why we employ categorical rules when analyzing regulations of speech. I do not mean to suggest that Professor Greenawalt is ultimately wrong with respect to his targeted vilification proposal; only that we must be very careful about adopting arguments that certain kinds of expression are less valuable than others.

Professor Greenawalt’s discussion of workplace harassment also reflects his values. Although this is an area which has been well-trod by others, n15 Professor Greenawalt starts by posing the right choice: either the workplace setting provides support for censorship which the government does not otherwise possess, or workplace harassment rules are inconsistent with fundamental [*322] First Amendment freedoms. Moreover, he does a fairly good job of going through the various arguments that support the proposition that the government can censor speech in the workplace that it otherwise cannot. He, quite properly I think, rejects arguments based upon the special nature of the workplace, a “captive audience” rationale, or the suggestion that restrictions of workplace speech constitute a “time, place or manner” restriction. n16 He spends time discussing the possibility that speech itself can sometimes discriminate, although I do not think this constitutes a particularly difficult or significant problem. n17

Ultimately, though, Professor Greenawalt relies primarily on another “low value” argument to support government regulation of workplace speech that is both crude and intended to demean or humiliate. Such speech – very much like his “targeted vilification” speech in discussing “hate speech” – is properly regulable, according to Greenawalt, with only a minimal showing of need by the government. But he goes further here and suggests that (1)repetitious, personally-directed speech in the face of a complaint by the recipient is punishable just as one might [*323] punish the “speech” of someone on the street who refuses to honor a request by someone else to be left alone and (2)even crude speech not intended to humiliate or intimidate (like crude sexual innuendo or the display of pornography) should be regulated on a showing of something less than “compelling need” because it is “low value” speech. n18 I pass over the first situation because Greenawalt is probably right in his analysis, and wrong in suggesting that such situations constitute any significant portion of workplace harassment claims based upon speech. n19

The second category (crude innuendo, pornographic calendars, and the like) warrants closer attention. Greenawalt very tentatively characterizes such speech as “low value,” n20 but it is far less clear than he suggests that the Supreme Court recognizes such a category. The Court occasionally has referred to the idea of low-value speech but, outside of the “fighting words” context, it has not upheld a content-based restriction of low-value speech, and it seems unlikely to do so after R.A.V. n21

More importantly, the new “low value” category Greenawalt suggests raises two additional problems. First, as I have already mentioned, using concepts like “low value” speech is a dangerous thing. We have a grave tendency to identify modes of communication that others use as “low value,” and the modes that we use as “high value.” n22 Expressing sexual desire in a [*324] crude way (as opposed to a polite, refined, and repressed way) may or may not be less effective (a point on which I express no opinion), but it is hardly less expressive. Indeed, one would think quite the opposite. When we suggest that one form of communication is better than another, we are simply imposing our own upper-middle-class, elite-school values on Constitutional doctrine. n23

Second, Professor Greenawalt has not really answered the question that he asked at the outset of his discussion on workplace harassment, viz., why should protection of speech in the workplace be any less than protection outside of the workplace? He rejects the traditional justifications for a difference in treatment that most advocates of workplace harassment rules use. He believes targeted harassment, either through repetition or through an intent to humiliate, should be without protection in both the workplace and on the street. To the extent he believes that a lower level of scrutiny should apply to other kinds of “low value” speech, he seems to rely on a (somewhat questionable) “low value” doctrine used by the Court outside of the workplace. His unstated conclusion should be, I think, the other possibility he mentioned at the outset of his discussion: that “the law of workplace harassment is out-of-kilter with general First Amendment doctrine.” n24 [*325]


In regulating speech among and between citizens, either inside or outside the workplace, the government acts in its role as sovereign. In the modern age of expansive government, though, government has many roles aside from that of treating citizens qua citizens. Should the role the government plays affect its ability to restrict speech? Should the government that can put people in jail be more constrained than the government that sells advertising space n25 or the government that operates public high schools or the government that operates public universities?

At least in some cases, the Supreme Court certainly thinks so and has so held when government as employer restricts the speech of its employees. n26 Indeed, the fact that a government employee receives any protection for his speech from the First Amendment represents a step up from accepted doctrine in the nineteenth century, which was that government employees “may have a constitutional right to talk politics, but [they] have no constitutional right to be [government employees].” n27 Today, to put the doctrine very roughly, a government employer can act in the same fashion that a typical and reasonable, viewpoint-neutral private employer would act. n28

Greenawalt quite properly notes that when a government promulgates a campus speech code for its state-run universities, it is acting in a different role than when it promulgates statutes against “hate speech.” Violation of rules against “hate speech” or “workplace harassment” can result in criminal penalties and civil damages, i.e., a restriction on one’s freedom or a taking of one’s property. Those who violate campus codes will simply have their status as students affected. Moreover, students at [*326] tending state-operated public universities receive significant state subsidies for their education. If a state wishes to promote the “civic virtues” or civil discourse, it may have a reasonable interest in seeing that state monies are not used to undermine those goals. n29

Thus, while it certainly cannot be said anymore that a student at a public university, who is punished for his speech, has a right to free speech, but no right to be a publicly-subsidized student, it does seem that the government’s role as an educator is different than its role as sovereign. Certainly, this is true at the secondary school level. n30 And while universities and colleges usually do not have the same concern with the civic and moral development of their students as high schools do, I am unaware of any Constitutional principle which precludes them from taking any interest. Greenawalt notes that “what is right for Columbia or Stanford may not be right for Brigham Young or Liberty Baptist College;” n31 but the same is true of public universities. What is right for the University of South Carolina may not be right for the Citadel. n32 The few lower court cases that have considered campus speech codes cited free speech precedents wherein the government was acting in its role as sovereign, and did not give any special consideration to the government’s role as educator. n33

Professor Greenawalt suggests that “courts should look not only to the concept of government as regulator of citizens but [*327] also to the concept of government as employer, in order to understand the proper relation of a state university to its students.” n34 I would put it more simply: perhaps the courts should try to understand the unique role of the state as educator, and the particular mission that a particular state institution is trying to accomplish. If my rough shorthand description of what a government qua employer can do is correct, perhaps government qua public universities should be allowed to restrict speech to the same extent that a reasonable and typical, viewpoint-neutral private university with a similar mission would. This standard constrains government regulation of speech more than the government-as-employer standard would, although it might change over time as universities do; but it would not place public universities at a grave disadvantage in effecting their “missions” viz a viz private universities. At this time in our history, moreover, this standard, I believe, is not all that different from the resolution that Greenawalt reaches, which would permit campus speech codes to preclude directed verbal harassment of other students, but require the protection of honest expressions of opinion no matter how offensive. n35

This would not save most modern campus speech codes. For the most part, they are not viewpoint-neutral and do not protect honest expressions of opinion. Moreover they are, in many instances, intolerably vague and overbroad. Indeed, perhaps the one common flaw in Greenawalt’s analysis of both workplace harassment and campus speech codes is his failure to recognize this, and his insistence that instances in which protected expression is implicated by these rules are rare and unusual. n36 They are not. Indeed, with both Title VII and Title IX now permitting compensatory and punitive damages for the victims of harassment (as well as permitting attorneys’ fees for the successful plaintiff; but not, in general, for the successful defendant), plaintiffs and their lawyers have every reason to push the harassment envelope. n37 Without any significant First Amendment

[*328] limit on the size of that envelope, I have little doubt that they

will. n38


In his penultimate chapter, Professor Greenawalt discusses how a consideration of “communities” might affect Constitutional analysis of various important First Amendment issues. Unfortunately, Greenawalt finds it difficult to define what a community is and never quite explains why judges (as opposed to, for example, legislators) should be weighing these values.

Greenawalt insists that there are communities smaller than “all of society” or the “polity” which are deserving of our consideration and attention (although he specifically refuses to focus upon whether local levels of government might be, or represent, a “community”). n39 He specifically cites cultural or religious groups that supply a strong sense of self-identification, like the Amish, but also wants to employ the term to include both units as small as families and collectives as large as “women.” n40 It is a “functional” definition, he tells us, but the independent variables of this particular function remain murky.

Once we find a “community,” what exactly should we do with it? If a men’s club could prove that it was a “community,” would that give it an extra claim to a right of association that permitted it to exclude women? Well, not exactly. Greenawalt concedes that “nongovernmental associations may be contexts of unfairness, domination, even coercion,” and agrees that intervention to rectify such unfairness “is often appropriate.” But, on the other hand, that intervention must be “carefully calibrated to the kind of association that is involved” and “sometimes the [*329] state must tolerate what most of its citizens deem to be unfair” in order to protect associational values. n41 Got that?

Things get no clearer when Greenawalt discusses specifics. The Court’s decision in Rust v. Sullivan, n42 which upheld Department of Health regulations precluding doctors who received federal funding for family planning services from advising patients of abortion services, threatened the “community” of doctor and patient. n43 (Greenawalt does not mention the community of those who do not want their money spent to support abortions.) Although veterans may be a “community,” their special interest in not having flags burned “should have little constitutional weight.” Most colleges and workplaces are not communities in a “narrow” sense and, in any event, Greenawalt concedes that the arguments about communities do not provide simple answers when discussing abusive speech since it is not altogether clear that rules against such speech really benefit all or most minority groups, because individual members of such groups have been known to employ “abusive” speech to make their points. n44

What Greenawalt does not discuss at all is the competence of the judiciary to analyze arguments about “communities.” Surely, the courts can accept rational value-judgments made by other branches about communities. But it is less than clear that when an argument about, for example, the “community” of doctor and patient has been given no weight by a co-equal branch of government, that the judiciary should step in to make that argument. The judiciary does have some institutional competence, we hope, in understanding individual rights and interpreting the provisions of positive law which give substance to those individual rights. But it frequently has trouble doing even this, and Greenawalt’s discussion of “communities” did not convince me that it is time for judges to branch out.

*** The criticisms set forth in this review aside, Professor Greenawalt has written a readable, succinct book on the First Amendment issues that are of most concern to Americans today. Even where I disagreed with, or could not penetrate, his analysis, I found his discussion of relevant cases informative and enlightening, and his ability to tie together seemingly disparate strands of [*330] First Amendment jurisprudence impressive. I plan to keep it handy as a ready reference.

THE CONSTITUTION AS POLITICAL STRUCTURE. By Martin Redish. n45 New York: Oxford University Press. 1995. Pp. 229. Hardcover, $ 39.95.