BANNING HATE SPEECH AND THE

STICKS AND STONES DEFENSE

Larry Alexander*

Black college students are the targets of racial epithets from

fellow students. Members of the American Nazi Party march
through the heavily Jewish Chicago suburb of Skokie, wearing
brownshirts and swastikas. Female employees encounter various
expressions of misogynistic or otherwise female-degrading views
from fellow employees or supervisors in the workplace. What
these events have in common is that in all of them, the targets of
the speech claim that the speech is harmful to them. Moreover,
in all of them, the targets have sought to have the hateful speech
legally suppressed through injunction (in Skokie),I campus hate
speech codes,2 or Title VII's ban on various types of job discrimi-
nation.3 Finally, in all of them, the speakers have raised a first
amendment defense against the attempts at legal suppression.
A lot of recent scholarly attention has been focused on these
and similar events, most of it concerned with whether and when
the first amendment should bar suppression of this "hate
speech."4 In addressing this issue, I, like most of the scholars,
shall take "hate speech" to mean epithets conventionally under-
stood to be insulting references to characteristics such as race,
gender, nationality, ethnicity, religion, and sexual preference.
Such epithets are the central targets of most campus speech
codes; and speech codes are unlikely to be drafted or interpreted
to apply to more measured expressions of views about race and

* Warren Distinguished Professor of Law, University of San Diego School of Law.

Visiting Professor of Law, University of Pennsylvania Law School, Fall, 1995.
I would like to thank participants in a U.S.D. School of Law workshop, Gail Heriot,
Fred Schauer, and Eugene Volokh for their helpful comments. I am particularly indebted
to Emily Sherwin, who commented extensively and perceptively on several drafts.
1. Collin v. Smith, 578 F.2d 1197 (7th Cir. 1978), cert. denied, 439 U.S. 916 (1978).
2. See, e.g., Doe v. University of Michigan, 721 F. Supp. 852 (E.D. Mich. 1989).
3. Civil Rights Act of 1964, 42 U.S.C. § 2000e (1991).
4. Mari J. Matsuda, et al., Words That Wound: Critical Race Theory, Assaultive
Speech, and the First Amendment (Westview Press, 1993); Henry Louis Gates, Jr., Let
Them Talk, New Republic 37 (Sept. 20 and 27, 1993); Kingsley R. Browne, Title VII as
Censorship: Hostile-Environment Harassment and the First Amendment, 52 Ohio St. L.J.
481 (1991).

71
72 CONSTITUTIONAL COMMENTARY [Vol. 13:71
the like, even if the expressed views are quite odious, because of
the obvious first amendment problems raised by any code ban-
ning more than epithets. Henry Louis Gates, Jr., has nicely illus-
trated the crucial distinction I wish to draw:
Contrast the following two statements addressed to a black
freshman at Stanford: (A) LeVon, if you find yourself strug-
gling in your classes here, you should realize it isn't your fault.
It's simply that you're the beneficiary of a disruptive policy of
affirmative action that places underqualified, underprepared
and often undertalented black students in demanding educa-
tional environments like this one. The policy's egalitarian
aims may be well-intentioned, but given the fact that aptitude
tests place African Americans almost a full standard deviation
below the mean, even controlling for socioeconomic dispari-
ties, they are also profoundly misguided. The truth is, you
probably don't belong here, and your college experience will
be a long downhill slide. (B) Out of my face, jungle bunny.s
The concern with hate speech is a concern with statements like
Gates's statement (B). Statement (A) is generally assumed to be
within the bounds of protected expression.6
It is generally accepted by those on both sides of this issue
that hate speech is harmful. The only question is whether,
notwithstanding its harmfulness, hate speech should be consid-
ered constitutionally protected free speech. It is my aim in this
brief essay to examine more closely the claim of harm. Exactly
how does hate speech harm its targets in paradigmatic contexts?
Among the questions I shall ask are whether it is really the
speech that is harmful, rather than what the speech reveals, and
whether, when it is the speech and not the revealed knowledge
that is harmful, the speech is harmful precisely for reasons that
should make it constitutionally protected. Finally, I shall ask
whether those who seek suppression of hate speech are really
seeking to be free of the harm it imposes or are instead more
ominously seeking suppression as an end in itself.
Before turning to the claim that hate speech is harmful, I
should say something about the contours of the constitutional
protection of speech that provides an essential background for
my analysis of the issue. Although there are several general the-
ories put forward by scholars and jurists to justify freedom of

5. Gates, New Republic (Sept 20 and 27, 1993) at 45 (cited in note 4).
6. I say "generally assumed" because there are several reported instances of legal
action taken against statements like statement (A) on the ground that such statements
constituted harassment. See Eugene Volokh, Freedom of Speech and Harassment, 39
UCLA L. Rev. 1791, 1800-07 (1992).
1996] BANNING HATE SPEECH 73

speech and its constitutional embodiment in the first amend-

ment, these theories generally converge on one central proposi-
tion. More importantly perhaps, this proposition has been
endorsed by the Supreme Court and represents settled first
amendment law. Put very simply, this proposition is that govern-
ment may not impose criminal or civil sanctions on speech be-
cause of its content, particularly because the content is, in the
government's view, wrong, dangerous, offensive, or disturbing.
As put by constitutional scholar Laurence Tribe, "[I]f the [first
amendment] guarantee means anything, it means that, ordinarily
at least, 'government has no power to restrict expression because
of its message, its ideas, its subject matter, or its content. .. .' "7
The general view reflected in both the theories of freedom of
speech and the Supreme Court's first amendment jurisprudence
is that it is better to let false or pernicious ideas compete in the
marketplace of ideas, where they are unlikely to prevail in the
long run, than to trust government to distinguish the false from
the true and the pernicious from the beneficial.
There are, of course, several exceptions to the first amend-
ment's proscription of governmental control of the content of
speech. Whether or not these exceptions are well-justified, what
is important for my purposes is that none of these exceptions is
relevant to the hate speech issue. For example, hate speech is not
usually an incitement to crime, a disclosure of a protected secret
or confidence, a violation of copyright, a defamatory injury to
reputation, the commission of fraud or perjury, or a misleading
commercial claim.s What I intend to show is that paradigmatic
hate speech, if harmful, is harmful only in ways that cannot jus-
tify its prohibition without gutting our constitutional conception
of freedom of speech. Whatever controversies surround first
amendment jurisprudence and its theoretical rationale do not
concern me, for here I operate squarely within the first amend-
ment's settled core. My interest lies with hate speech and the
harms attributed to it, not with the first amendment.

I. THE CLAIMS OF HARM

Hate speech is alleged to be harmful in several different
ways. First, it is insulting, and insults are psychologically wound-

7. Laurence H. Tribe, American Constitutional Law 790 (The Foundation Press,

Inc., 2d ed. 1988), 790 (quoting Police Department of Chicago v. Mosley, 408 U.S. 92, 95-
96 (1972)).
8. I discuss the relevance of the "fighting words" exception at notes 26-27, infra.
And I discuss speech by public employees, including teachers, at notes 74-79 infra.
74 CONSTITUTIONAL COMMENTARY [Vol. 13:71
ing and cause emotional distress.9 Second, it creates unequal op-
portunity in the school and workplace [Link] Third, it
silences those who are its targets, depriving them of their free-
dom of speech.u Fourth, it offends by flouting social norms re-
garding proper verbal behavior.12 And, fifth, its expression is a
speech act that shows disrespect for or even subordinates its
targets.13
There is no gainsaying the first type of harm hate speech
causes, the harm of psychological pain. What is interesting about
this type of harm is not its existence but its connection to the hate
speech that causes it. I shall examine this connection in the next
section.
The next two types of harm-unequal opportunity and loss
of freedom of speech-are much more problematic. They would
be unproblematic if the claim on their behalf were merely that
those who suffer insults are less likely than others to perform up
to their potential or to participate in the public exchange of
ideas. That claim may well be true, but it would add little to the
claim that we should protect people from the harm of insult. All
of the discussion of unequal opportunity and "silencing" of
speech would just be portentous ways of talking about and cash-
ing out the psychological harm caused by insults.
Perhaps, however, there is something more to the claims of
unequal opportunity and silencing of speech. At least some-
times, the claim is that the speech will cause others who hear it or
hear of it to lower their opinions of the targets and become more
dismissive of their ideas, their accomplishments, and their needs.
The claims of unequal opportunity and silencing translate into
the claim that an insult of the target, even if only an epithet, car-
ries a propositional content for an audience-for example, "the
target and those like her are unworthy of respect" -which pro-
positional content might be accepted as true by the audience to
the detriment of the target's status as a student, an employee,
and a participant in public dialogue.
This last claim regarding the harm of hate speech is insuffi-
cient to justify withholding first amendment protection and, in-
deed, standing alone, helps to make the case for first amendment

9. Matsuda, Words that Wound at 67-68, 90-96 (cited in note 4).

10. See, e.g., 29 C.F.R § [Link](a)(3) (1995).
11. Matsuda, Words that Wound at 78-79 (cited in note 4).
12. Joel Feinberg, 2 Moral Limits of the Criminal Law, Offense to Others 16-18 (Ox-
ford U. Press, 1985).
13. Andrew Altman, Liberalism and Campus Hate Speech: A Philosophical Exami-
nation, 103 Ethics 302, 309-12 (1993).
1996] BANNING HATE SPEECH 75

protection. The fact that an opinion about others (1) is wrong

and (2) might persuade others cannot in general be grounds for
withholding first amendment protection without turning the first
amendment on its head.14 So I shall ignore this claim and focus
on the claim that hate speech is directly harmful to its targets
rather than indirectly harmful through its effects on others' atti-
tudes and beliefs about the targets. That is, I shall deal with the
alleged harms of emotional distress, offense, and subordination.
My approach to this subject is organized as follows. In Part
II, I focus on the various kinds of painful knowledge that hate
speech conveys, both denotatively and connotatively, and the
various ways such knowledge can be painful. My conclusion,
however, is that hate speech conveys nothing that cannot be con-
veyed equally or more painfully by other speech, speech that
clearly is and should be constitutionally protected. Insofar as
painful knowledge is concerned, there is nothing distinctive
about hate speech.
In Part III, I tum my attention to the interplay between hate
speech and social norms, norms that determine what gives of-
fense and what counts as the performance of various types of
speech acts, such as acts of subordination and disrespect. I con-
clude that such social norms should not limit freedom of speech,
and that civility in public discourse, however desirable, should
not be legislated.
In Part IV, I take up some possible responses to my analysis,
in particular, arguments for carving out an exception from first
amendment protection either for epithets or for the ideas ex-
pressed by epithets. In responding to those arguments I offer
some general observations about the difficulty of making out a
case for freedom of speech protection on non-deontological
grounds.
In Part V, I make a concession to those who would ban hate
speech, namely, that the Constitution probably permits them to
do so in state colleges and universities. Finally, in Part VI, I spec-
ulate about what lies behind the recent moves to ban hate
speech.

14. See text at note 7 supra.

 76 CONSTITUTIONAL COMMENTARY [Vol. 13:71
II. INTRINSICALLY HARMFUL SPEECH AND
PROTECTING IGNORANCE
Michael Sandel recently characterized hate speech of the
type I am considering as intrinsically [Link] If the harm to
which Sandel is referring is the psychological distress of having
been insulted, then it surely cannot be correct that the speech is
intrinsically harmful qua hate speech. Let me illustrate why with
an example that should work well for all the contexts of hate
speech that we are considering.
Suppose a white employee mutters "damn nigger" whenever
he encounters a black employee, loudly enough to be heard by
the black employee. Black employees complain to the supervi-
sor, who fails to take any action against the white employee, per-
haps because the supervisor shares the white employee's
sentiments. The black employees sue under Title VII alleging a
racially hostile work environment and thus illegal racial
discrimination.
Now the hate speech in question might well be quite psycho-
logically painful, but is it intrinsically harmful qua hate speech?
Are the words uttered "words which by their very utterance in-
flict injury"?
I say no. What is psychologically painful in this case is
knowledge, specifically knowledge of the white employee's (and
perhaps the supervisor's) attitudes and beliefs.t6 The hate speech

15. Remarks at the annual meeting of the American Association of Law Schools,
Section on Jurisprudence, January, 1993.
16. Gates makes a similar point in response to the contention that epithets harm in
some way other than by conveying harmful ideas:
Nor, finally, does the Chaplinsky-derived description of assaultive speech as
being devoid of political or other ideational content-"experienced as a blow,
not a proffered idea," in [Charles] Lawrence's compelling formulation-survive
closer inspection. Consider the incident that, Lawrence tells us, moved him to
take up the hate speech cause in the first place. 1\vo white Stanford freshmen
had an argument with a black student about Beethoven's ancestry: he claimed,
and they denied, that the Flemish-German composer was really of African de-
scent. The next evening, apparently as a satirical commentary, the white stu-
dents acquired a poster of Beethoven, colored it in with Sambo-like features,
and posted it on the door of the student's dorm room at Ujamaa, Stanford's
black theme house. Lawrence "experienced the defacement as representative of
the university community's racism and not as an exceptional incident in a com-
munity in which the absence of racism is the rule"-and the rest is critical-race-
theory history.
Now then, is Lawrence's paradigm example of racist hate speech in fact
devoid of ideational or political content, as his analysis would suggest? Evi-
dently not, for in their jointly written manifesto for critical race theory, the au-
thors of Words That Wound spell out what they believe its message to have been:
"The message said, 'This is you. This is you and all of your African-American
brothers and sisters. You are all Sambos. It's a joke to think that you could ever
be a Beethoven. It's ridiculous to believe that you could ever be anything other
1996] BANNING HATE SPEECH 77
is quite good evidence of those attitudes and beliefs, but knowl-
edge of those attitudes and beliefs might have been obtained
many other ways. Some of those ways might not involve the su-
pervisor's speech but rather his actions (which, of course, do
"speak" in the sense of communicating information about his at-
titudes and beliefs). The black employee might notice Ku Klux
Klan literature on the white employee's desk. Or the white em-
ployee might be observed after work wearing a Ku Klux Klan tie
clasp.
Moreover, even if the white employee's speech is the source
of the black employee's knowledge of the former's hateful atti-
tudes and beliefs, the speech might be speech that (almost)
everyone, and surely the Supreme Court, would treat as pro-
tected by the first amendment. The white employee might take
out an ad in the newspaper stating, "Were it not for Title VII, I
would utter racial epithets to black employees." That ad would
convey exactly the same painful knowledge to the black em-
ployee as the racial epithets in the initial example. So, too,
would a sign over the white employee's desk that said the same
thing.n
Moreover, if one already possesses the painful knowledge
that another thinks ill of members of one's race-so ill that he
would hurl racial epithets at them-his actually doing so would
not itself be additionally painful to any great degree. After a
black observer listens for an hour to hate-mongering speeches at
a Ku Klux Klan rally, all of which speeches are protected by the

than a caricature of real genius.' " The defaced poster would also inspire a
lengthy and passionate essay by the legal theorist Patricia J. Williams, an essay
that extracts an even more elaborated account of its meaning. This was one
picture, clearly, that really was worth a thousand words.
The same paradox surfaces in Richard Delgado's ground-breaking proposal
for a tort action to redress racist speech. . . . Delgado argues that a racial insult
" ... is not political speech; its perpetrator intends not to discover truth or advo-
cate social action, but to injure the victim." It's a curious disjunction, this, be-
tween advocacy and injury. For if Delgado and his fellow contributors have a
central message to impart, it's that racial insults are profoundly political, part of
a larger mechanism of social subordination, and thus in contravention of the
spirit of the "equal protection" clause of the Fourteenth Amendment. And the
most harmful forms of racist speech are precisely those that combine injury with
advocacy -those that are, in short, the most "political."
"Are racial insults ideas?" Lawrence asks. "Do they encourage wide-open
debate?" He means the question to be rhetorical, but after reading his work and
those of his fellow critical race theorists, who could possibly doubt it?
Henry Louis Gates, Jr., War of Words: Critical Race Theory and the First Amendment, in
Henry Louis Gates, Jr., et al, Speaking of Race, Speaking of Sex 17, 25-27 (New York U.
Press, 1994).
17. Although my assumption is that most courts would treat such indicators of em-
ployee beliefs as constitutionally protected expression-as they should-not all courts
have done so. See Volokh, 39 UCLA L. Rev. at 1800-07 (cited in note 6).
 78 CONSTITUTIONAL COMMENTARY [Vol. 13:71

first amendment, would he be additionally pained if one of the

Klan, upon seeing him, uttered a racial epithet?ts
There is, admittedly, a common intuition that there is some-
thing special-and especially painful-about hate speech, and
this intuition is widespread enough to require explanation. Hate
speech seems different from other denotatively comparable
speech because it carries a particular affective tone, a tone that
adds vehemence to the insult it conveys. For this reason, hate
speech may inflict a special sort of pain: the hearer learns not
only that the speaker is biased against persons of his type, but
also that the speaker's dislike is so intense that he has chosen to
express it in terms conventionally associated with insult.
Does this difference in affective tone identify a distinction
between hate speech and other negative expressions about race
and similar characteristics? The pain it causes still is a product of
the hearer's knowledge of the speaker's views, though the knowl-
edge may differ somewhat from that conveyed by another form
of expression. Moreover, this new increment of knowledge-be-
yond or different from that conveyed by more "civil" expres-
sions-hardly explains the indignation with which opponents of
hate speech have pursued the cause of banning it. One suspects
that they are reacting, not to the extra vehemence attached to
epithets, but to the underlying racial, sexual, or other bias.
More importantly, the same sort of pure and vehement dis-
dain conveyed by hate speech can be achieved easily enough in a
different form. Just as the affective tone that epithets usually
carry can be absent in some uses of them, so too can that same
affective tone be conveyed as or more efficiently through superfi-
cially polite speech. Consider the examples offered by Gates.19
Surely Statement (A), the polite putdown, can be much more
devastating than Statement (B), the use of a racial epithet. That
is Gates's point, and he is correct. Banning epithets, but not
skillful rhetorical skewerings, would essentially and unjustifiably
discriminate against low-brow forms of expression. Hate speech
may convey painful knowledge in a particularly painful way, but
it is not unique in that ability.
My conclusion is that what people may wish to avoid when
they seek to ban hate speech in contexts like Skokie, campuses,

18. I am assuming that the utterance does not occur in a context that would bring it
within the special classes of fighting words or assaults, or that would cause in the listener
some seriously distracting urge to reply. See text accompanying notes 26-29 and text at
pp. 81-82 infra.
19. See text at notes 5-6 supra.
 1996] BANNING HATE SPEECH 79

and the workplace is knowledge: knowledge that the hate speech

is sufficient but often unnecessary to convey. The knowledge op-
ponents wish to avoid is undoubtedly painful. But should people
be given legal entitlements to be left in ignorance of others' atti-
tudes towards and beliefs about them when those entitlements
entail legal prohibitions of expression?
One might disagree with my hypothesis that knowledge of
others' attitudes and beliefs is what is at issue by arguing that it is
really the hateful attitudes and beliefs themselves, not knowledge
of them, that people wish to be spared. But if that were true, it
would surely destroy the case for banning hate speech: if there is
anything the first amendment precludes, it is attempts at thought
contro1,2o (Additionally, banning hate speech often leaves the
hateful attitudes and beliefs intact and just creates the illusion
that they have disappeared.)
Similarly, if one argues that the harm of hate speech is not
the painful knowledge it gives its target but rather the persuasive
effect it may have on the audience beyond the target, then, as I
said in the first section, that is a conclusive reason under the first
amendment for its protection, not its suppression.z1
So knowledge, and the desire not to have it, appears to be a
plausible target of the move to ban hate speech. Can a case be
made that, notwithstanding the first amendment, the state can
punish or impose liability on speakers because others do not wish
to know the speakers' attitudes and beliefs?

20. One of the earliest academic calls for sanctions against epithets in fact rests its
case as much on a thought-control rationale as on the claim that epithets are directly
injurious. See Richard Delgado, Words That Wound: A Ton Action for Racial Insults,
Epithets, and Name-Calling, 17 Harv. Civ. Rts.-Civ. Lib. L. Rev. 133 (1982):
The racial insult remains one of the most pervasive channels through which dis-
criminatory attitudes are imparted. . . . Not only does the listener learn and
internalize the messages contained in racial insults, these messages color our
society's institutions and are transmitted to succeeding generations.
ld. at 135-36 (footnotes omitted).
***
The establishment of a legal norm "creates a public conscience and a standard
for expected behavior that check overt signs of prejudice." Legislation aims first
at controlling only the acts that express undesired attitudes. But "when expres-
sion changes, thoughts too in the long run are likely to fall into line."
ld. at 149 (footnotes omitted).
21. I am obviously assuming that such a ban could not survive invalidation under the
very narrow exception the Supreme Court has carved out for proscription of libel. See
New York Times v. Sullivan, 376 U.S. 254 (1964). But see Beauharnais v. Illinois, 343 U.S.
250 (1952). Henry Louis Gates, Jr. argues effectively that racial invective is completely
disanalogous to individual libel: "You cannot libel someone by saying 'I despise you,'
which seems to be the essential message common to most racial epithets." Gates, Speak-
ing of Race at 30 (cited in note 16).
80 CONSTITUTIONAL COMMENTARY [Vol. 13:71
There are some easy cases where the answer is surely yes,
and others where the answer is surely no. For example, we have
zones of privacy, such as inside our homes, where we can demand
that certain ideas not be communicated to us and enlist the aid of
the state in excluding them (for example, commercial or charita-
ble solicitations, or obscenity).22 On the other hand, a public
speech urging a willing audience to vote Libertarian, delivered in
a public forum or in a newspaper, cannot be legally suppressed
even if the attitudes and beliefs conveyed are psychologically
painful to bystanders.
The contexts in which the attempts to ban hate speech arise
are contexts that are not private, even if they are not always par-
adigmatic public fora. Skokie did involve a paradigmatic public
forum.23 And campus hate speech codes frequently cover and
are applied to epithets uttered anywhere on campus. The work-
place is not a traditional public forum because the employer can
limit speech there to prevent disruption of production and per-
haps even to further the employer's own ideological ends.24 Yet
the hate speech issue arises just when the employer has not
banned speech in the workplace and is allowing speech that is
not incompatible with the assigned tasks, and then it is the em-
ployee, not the employer, who is seeking to ban the speech. Put
differently, in all the contexts with which I am dealing, one must
look at the issue as one of public forum speech in the sense that
the fora are surely not within the target's conceded zone of
privacy.2s

22. Rowan v. United States Post Office, 397 U.S. 728 (1970).
23. The Skokie case involved denial of pennission to march on city streets. Collin,
578 F.2d at 1197.
24. See Stanley Ingber, Rediscovering the Communal Worth of Individual Rights:
The First Amendment in Institutional Contexts, 69 Tex. L. Rev. 1 (1990).
25. Kingsley Browne makes what I believe is an effective response to the claim that
the workplace is not a public forum in situations where the employer is not restricting
employee speech:
Robert Post has suggested a somewhat different justification for limiting
speech in the workplace. Although he argues that limitations on racist speech in
"public discourse" are highly suspect, he asserts that speech in the workplace
does not generally constitute public discourse. Post defines "public discourse"
as "encompassing the communicative processes necessary for the formation of
public opinion, whether or not that opinion is directed toward specific govern-
mental personnel, decisions, or policies. However, asserts Post, "within the
workplace ... an image of dialogue among autonomous self-governing citizens
would be patently out of place."
Post's argument seems to presuppose that the communication contributing
to public opinion is largely limited to the press, handbillers on public streets, and
fiery orators in the parks. Yet, for most citizens-who are not political activ-
ists-the great bulk of their discussion of political and social issues probably
occur [sic] in the home and the workplace. For example, there are probably very
few workers in the United States who did not discuss the Gulf War while at
1996] BANNING HATE SPEECH 81

If one assumes, then, no zone of privacy rationale that would

support a ban on any speech to which those protected by that
zone objects, one must ask whether there are any other grounds
supporting bans on hate speech because of the psychological pain
it inflicts.
I shall leave aside the "fighting words" doctrine.z6 That doc-
trine has been given little scope by the Supreme Court,27 and in
any event it covers only a limited number of the instances of hate
speech at issue. Moreover, my inquiry is whether the interests of
the targets of hate speech support its suppression, for it is primar-
ily they who are urging this. The fighting words exception to
freedom of speech is designed primarily to allow the public to
protect against breaches of the peace and not out of concern for
those provoked to illegal violence.
I shall also leave aside those rare instances in which uttering
epithets constitutes a criminal assault-a threat to commit an im-
minent battery. Epithets may be employed to threaten, but again

work. Presumably, Post would argue for protection of those discussions under
the first amendment, and I would hope he would do so even if the discussions
offended workers having Iraqi citizenship. If he would protect those discussions,
it is difficult to understand why he would withdraw similar protection for speech
that conveys ideas offensive to women and minorities.
Post's suggestion that workplace speech is not public discourse places too
much emphasis on where the speech takes place and too little emphasis on the
content of the speech. As Post acknowledges, "[s]peech that can be said to be
about matters of 'public concern' is ordinarily classified as public discourse." It
should not lose the protection of the first amendment simply because someone
chose to express it on private property.
In sum, the term "workplace" is not a talisman that extinguishes first
amendment protections. Just as school children "do not shed their constitu-
tional rights to freedom of speech or expression at the schoolhouse gate," work-
ers do not shed theirs at the factory gate. The Supreme Court has narrowly
limited the circumstances in which expression in the workplace may be regu-
lated, and those circumstances have no relevance to hostile-environment harass-
ment cases.
Browne, 52 Ohio St. L.J. at 515-16 (cited in note 4) (footnotes omitted).
Richard Fallon also suggests that epithets in the workplace are uniquely harmful,
even if they do no more than convey painful knowledge of others' attitudes. Unfortu-
nately, Fallon gives no argument for why this is so. Moreover, the argument he constructs
against his position-that one should in general have no right to be protected from
knowledge of others' opinions-seems to me persuasive. See Richard H. Fallon, Jr., Two
Senses of Autonomy, 46 Stan. L. Rev. 875, 897-98 n.138 (1994).
26. Chaplinsky v. New Hampshire, 315 U.S. 568 (1942).
27. City of Houston v. Hill, 482 U.S. 451 (1987); Eaton v. City of Tulsa, 415 U.S. 697
(1974); Lewis v. City of New Orleans, 415 U.S. 130 (1974); Gooding v. Wilron, 405 U.S.
518 (1972). Many commentators believe that the fighting words doctrine has disappeared
as a separate exception to the general constitutional proscription of content regulation,
and that only those fighting words which also count as incitements to imminent lawless
action remain punishable. See Brandenburg v. Ohio, 395 U.S. 444 (1969); Tribe, Ameri-
can Constitutional Law at 850-51 (cited in note 7); Nadine Strossen, Regulating Racist
Speech on Campus: A Modest Proposal?, 1990 Duke L.J. 484, 510-11. See also Gates,
Speaking of Race at 25 (cited in note 16).
 82 CONSTITUTIONAL COMMENTARY [Vol. 13:71
they are not unique in this regard. A ban on hate speech qua
epithets is redundant of existing criminal statutes if its object is
criminal asssaults; and as a special proscription of one form of
criminal assault, a hate speech ban is probably unconstitutionally
discriminatory.2s In any event, my interest lies with bans on epi-
thets that do not meet the stringent criteria for criminal assaults
unprotected by the first amendment.29
Having excluded these recognized but narrow categories of
unprotected speech and narrowed the focus, let me return to the
principal question regarding hate speech, namely, whether and
when speech can be suppressed to protect others from knowl-
edge that they find painful.
One situation in which the claimed right to be spared painful
knowledge surely cannot override freedom of speech is the situa-
tion in which the speech is directed at an audience other than
those who are pained by it. If an employee (E) overhears speech
between fellow employees or bosses that reveals negative atti-
tudes towards and beliefs about E or those of E's race or gender,
E should have no right to have the speech legally suppressed or
punished or made the basis of civilliability.3o It does not make
any difference where the speech occurs-in a newspaper, on a
talk show, at a political rally, or in the workplace-so long as it
does not occur within E's zone of privacy. Nor does it matter
whether the speech is scholarly or consists primarily of epithets.
It is the knowledge about beliefs and attitudes that is paining E,
and the words are just evidence of the beliefs and attitudes.
What about speech that is directed, not at another audience,
but at the target of hostility, that is, at E himself? E's concern
here is, I am assuming, avoiding painful knowledge, not keeping
others who might overhear the speech or learn of it from being
persuaded themselves to hold similar attitudes and beliefs. And
to strengthen E's case even further, I shall assume that E has in
no way indicated that he seeks the knowledge in question.
In these circumstances-no other audience and no solicita-
tion of the speech-can government suppress speech to protect
E from information E finds painful? What would support such a
power beyond the mere fact that the information is painful?

28. See R.A. V. v. City of St. Paul, 112 S. Ct. 2538 (1992).
29. See generally Kent Greenawalt, Speech, Crime, and the Uses of Language 90-91
(Oxford U. Press, 1989).
30. But see Delgado, 17 Harv. Civ. Rts.-Civ. Lib. L. Rev. at 135-36, 149 (cited in
note 20).
 1996] BANNING HATE SPEECH 83
First, E might claim that the speech is purely malicious. If
there is no other audience for it, and he himself has not sought it
and instead wishes to avoid it, then it can serve no purpose ex-
cept to inflict pain on him. Surely the first amendment should
not protect speech of that sort.31
One problem with this line of attack on hate speech is that
although a lot of hate speech is undoubtedly motivated by the
desire to inflict pain, a lot of it is entirely or partially the product
of other motives. For instance, a fellow employee might wish to
tell E how much he hates E and those like E in order to ward off
E's attempts to befriend him, to share confidences with him, and
so forth.32
Moreover, it is hard to see how a right against malicious
painful speech could be limited to epithets. Once one under-
stands that epithets are painful, not intrinsically, but because of
the beliefs and attitudes they reveal, which beliefs and attitudes
can be revealed by all sorts of speech and actions, then a right
against painful malicious speech translates into a right not to be
told in any form about unwelcome beliefs and attitudes if the
speaker intends to inflict psychological pain. In short, racial and
sexual epithets just become instances within a broad category of
malicious infliction of emotional distress.33
Although there is a generally recognized tort of infliction of
emotional distress, which, if constitutionally valid, might support
criminal or civil penalties for speech inflicting emotional dis-
tress,34 the Supreme Court has indicated that it regards that tort
as constitutionally suspect when applied to speech, or at least
speech that does not consist of false statements of fact. In Hus-
tler Magazine v. Falwel/,35 the Reverend Jerry Falwell sued Hus-
tler Magazine for printing a particularly tasteless spoof ad in
which Falwell admits, among other things, to having sex with his
mother. One of Falwell's claims against Hustler was for inten-

31. See Browne, 52 Ohio St. LJ. at 546 (cited in note 4).
32. See Alan E. Brownstein, Hate Speech and Harassment: The Constitutionality of
Campus Codes That Prohibit Racial Insults, 3 Wm. & Mary Bill of Rights J. 179, 204-5
(1994); Kent Greenawalt, Insults and Epithets: Are They Protected Speech?, 42 Rutgers L.
Rev. 2f57, 298 (1990) (separating intent to harm from an honest but hurtful statement is
difficult).
33. See Greenawalt, 42 Rutgers L. Rev. at 300 (cited in note 32) (noting the line-
drawing problems with a cause of action for intentional inftiction of emotional distress
based on epithets).
34. See American Law Institute, Restatement (Second) of Torts § 46 at 71 (Ameri-
can Law Institute Publishers, 1965): "One who by extreme and outrageous conduct inten-
tionally or recklessly causes severe emotional distress to another is subject to liability for
such emotional distress .... "
35. 485 u.s. 46 (1988).
84 CONSTITUTIONAL COMMENTARY [Vol. 13:71
tional infliction of emotional distress. The Supreme Court re-
versed the jury's verdict for Falwell on first amendment grounds
and held that Falwell could not recover damages for infliction of
emotional distress without showing that the defendant had made
a false statement of fact with knowledge or reckless disregard of
its falsity.36 Robert Post interprets the Court's opinion in Falwell
as granting first amendment protection to outrageous expres-
sions that are part of public discourse, no matter how painful,
and no matter the motivation.37 Because the Supreme Court is
unlikely to deem hate speech to lie outside public discourse or to
be false statements of fact,38 it is unlikely to find either outright
bans of hate speech or tort claims based on hate speech to be
constitutionally valid methods of protecting against emotional
distress.39 Moreover, as I have already shown, there is no emo-
tional distress that hate speech can inflict that cannot be inflicted
as well or better by "polite" discourse.4o
Second, E might claim, not that the speech is purely mali-
cious, but that the speech is injuriously distracting. How might
hate speech be so? Well, some speech reveals information that
remains in the forefront of the mind even when tasks at hand
require that the information be relegated to the subconscious so
that the mind can concentrate on other things. This can, of
course, be true of hate speech, especially if it is frequent or, as in,
for example, the case of a permanent sign, constant. The knowl-
edge hate speech brings about will be not only painful but also
distracting. And the distraction might in tum interfere with per-
formance as an employee or as a student.

36. Falwell, 485 U.S. at 56.

37. Robert C. Post, The Constitutional Concept of Public Discourse: Outrageous
Opinion, Democratic Deliberation and Hustler Magazine v. Falwell, 103 Harv. L. Rev.
603, 612-14 (1990).
38. See Robert C. Post, Racist Speech, Democracy, and the First Amendment, 32
Wm. & Mary L. Rev. 267 (1991).
39. See American Law Institute, Restatement (Second) of Torts § 46 comment d
(American Law Institute Publishers, 1965). Comment d states in part:
The liability clearly does not extend to mere insults, indignities, threats, an-
noyances, petty oppressions, or other trivialities. The rough edges of our society
are still in need of a good deal of filing down, and in the meantime plaintiffs
must necessarily be expected and required to be hardened to a certain amount
of rough language, and to occasional acts that are definitely inconsiderate and
unkind. There is no occasion for the law to intervene in every case where some
one's feelings are hurt. There must still be freedom t~ e~pre~ an unflattering
opinion, and some safety valve must be left through wh1ch 1r8SC!ble tempers may
blow off relatively harmless steam.
40. Of course, where the speech is directed at E, and there is no other audience for
it, the speech is arguably not part of "public discourse." In such a situation, however, the
speech is unlikely to be severely emotionally distressing. Jerry Falwell would undoubt-
edly have been much less distressed by being insulted privately by Hustler's publisher.
1996] BANNING HATE SPEECH 85

There is another way that hate speech might be distracting,

and that is by provoking responses. For example, E might not
even be pained or distracted by the knowledge that the supervi-
sor despises E's race or gender, or E might have gained that
knowledge by means other than the supervisor's epithets. But
epithets, when directed at E, might, in conjunction with social
conventions (or perhaps even biological hard wiring), create in E
a felt need to respond to the supervisor. And responding-or
resisting the urge to respond-could be sufficiently distracting to
interfere with E's performance.
Again, both arguments from distraction are applicable to
much more than hate speech primarily and narrowly conceived
as racial, sexual, and other group-regarding epithets. A lot of
speech reveals information that is difficult to banish from con-
sciousness and that, if it remains in the forefront of the mind,
might interfere with job or educational performance. And a lot
of speech might create an urge to respond, including speech aver-
ring propositions that the listener regards to be in error, even if
the speech is not in any way about the listener. A right to ban
speech that is distracting because of its content could be very
broad indeed, even if limited to cases where there is no third
party audience.

III. SOCIAL NORMS AND SPEECH ACTS: THE HARMS

OF OFFENSE, SUBORDINATION, AND
DISRESPECT
Thus far I have dismissed the claims that hate speech by its
very nature creates unequal opportunity or "silences" its targets.
I have focussed at some length on the claim that hate speech
causes psychological harm, a claim that I accept. The problems
with banning hate speech for this reason are that such a ban
would rest on the dubious ground that speech can be banned to
protect others from unwanted knowledge-a ground that is ac-
ceptable at most only within certain zones of privacy-and that a
ban so justified extends logically far beyond the sort of epithets
normally thought of as hate speech.
There are two other possible harms hate speech might cause.
Hate speech-or more specifically, the epithets it characteristi-
cally employs-may violate social norms against using such epi-
thets, which in tum may give rise to offense. And hate speech
might be viewed as a speech act that counts as disrespecting or
even socially subordinating its targets.
86 CONSTITUTIONAL COMMENTARY [Vol. 13:71

A. HATE SPEECH AND OFFENSE

I shall not quarrel with the contention that the epithets of

hate speech flout social conventions and thereby give rise to of-
fense. My quarrel is with allowing offense to suffice as a ground
for suppressing speech. Some philosophers, such as Judith
Thomson, deny that one has any rights against infliction of belief-
mediated harms.4t Thus, Thomson would object to using the psy-
chological distress of insult that I discussed in the previous sec-
tion as a reason for banning speech (or any other behavior), even
when the speech is malicious. Likewise, she would object to any
putative right not to be offended. Even if one does not go as far
as Thomson in rejecting rights not to be subject to belief-medi-
ated distress-and I do not-one should be very wary of using
offense as a ground for restricting freedom of speech.
Moreover, the Supreme Court itself has been quite adamant
that offense is not a legitimate reason for suppressing speech ex-
cept when the privacy of the home is invaded.42 For example, in
Cohen v. California,43 the Court reversed Cohen's conviction for
breach of the peace despite the fact that Cohen had surely
flouted norms in an offense-generating way by parading in public
wearing a jacket emblazoned with "Fuck the draft." Similarly, in
Erznoznik v. City of Jacksonville,44 the Court refused to allow
the city to ban movies with nude scenes in drive-in theaters with
screens visible to passing motorists.
One does not have to deny the reality or the desirability of
the social norms whose flouting causes offense to maintain that
potential offense should not bar speech in the public domain. As
the Supreme Court has recognized, the line between an offensive
word and an offensive idea is a line too difficult to draw to be
serviceable to the first amendment.4s

41. Judith Jarvis Thomson, The Realm of Rights 253-59 (Harvard U. Press, 1990).
42. See F.C.C. v. Pacifica Foundation, 438 U.S. 726 (1978).
43. 403 u.s. 15 (1971).
44. 422 u.s. 205 (1975).
45. See Greenawalt, 42 Rutgers L. Rev. at 300 (cited in note 32) (presenting general
offensiveness as a basis for restricting hate speech); Strossen, 1990 Duke L.J. at 549 (cited
in note 27) (arguing that racial insults convey racial ideas). But see Delgado, 17 Harv.
Civ. Rts.-Civ. Lib. L. Rev. at 175 (cited in note 20) (attempting to distinguish Cohen).
The phenomenology of offense is complex. Sometimes, the expression that causes
the offense does so by focusing one's attention on unpleasant or disgusting ideas. In such
cases, offense is similar to unwanted distraction.
At other times, offensive expressions cause embarrassment, as, for example, when
they are uttered in one's presence when one is accompanied by another (say, one's
spouse, parent, or child). In such cases, what is embarrassing is being forced to think
about something that you also know your companion, present when the offensive com-
ment is made, knows you are now thinking about. Epithets cause a similar psychological
1996] BANNING HATE SPEECH 87

B. HATE SPEECH AS A SuBoRDINATING SPEECH Acr

Andrew Altman in a recent article appears to agree with the
points I have made thus far.46 He points out with respect to the
various psychological harms that hate speech inflicts that banning
hate speech because of these harms "sweeps too broadly for a
liberal to countenance it. Forms of racist, sexist, or homophobic
speech that the liberal is committed to protecting may cause pre-
cisely the kinds of harms that the proposed justification in-
vokes. "47 Altman identifies "racist, sexist, or homophobic
speech couched in a scientific, religious, philosophical, or polit-
ical mode of discourse" as speech that causes those kinds of
harms but that surely must be protected.48
Altman nonetheless supports banning some hate speech,
that which counts as subordinating speech acts. Altman distin-
guishes between hate speech's perlocutionary effects-its causal
effects on the audience, which in the case of hate speech do not
support its suppression-and hate speech's illocutionary force,
which Altman contends amounts to the wrong of treating a per-
son as having inferior moral standing. "[H]ate speech involves
the performance of a certain kind of illocutionary act, namely,
the act of treating someone as a moral subordinate."49
Altman contends that the conventional rules of language
distinguish between, for example, arguing that blacks are geneti-
cally inferior to whites, which is not itself the performance of an
act treating blacks as moral inferiors, and calling blacks "nig-
discomfort because, when directed at one, one must struggle with resisting the distraction
of a reply, a struggle that those who uttered the epithet are aware is taking place. (A
similar phenomenon explains the painfulness of unwanted sexual remarks directed at wo-
men in public places. See Cynthia Grant Bowman, Street Harassment and the Informal
Ghettoization of Women, 106 Harv. L. Rev. 517 (1993).)
As with offensive expression, epithets not only inflict unpleasant thoughts, embar-
rassment, and the urge to reply, but also in so doing flout social norms of civility, giving
rise to feelings of indignation. And as with offensive comments, epithets frequently carry
an affective tone different from other expressions with which they are literally
synonymous.
Nonetheless, forbidding specific words because they are offensive epithets cannot
prevent any of these harms from occurring. As I have pointed out, a very skillful rhetori-
cal skewering of someone can outdo epithets in causing unpleasant, embarrassing, or dis-
tracting thoughts, in giving rise to indignation by flouting norms of civility, and in evoking
a visceral response through its affective tone. See, e.g., the text at note 19 supra. And as I
have also previously argued, banning epithets, but not the skillful rhetorical skewerings,
essentially discriminates against low brow forms of expression.
For an excellent account of the phenomenology of offense, see Feinberg, Offense to
Others at 10-22 (cited in note 12).
46. Altman, 103 Ethics 302 (cited in note 13).
47. Id. at 306.
48. Id.
49. Id. at 309-10.
88 CONSTITUTIONAL COMMENTARY [Vol. 13:71
gers," which [Link] And speech acts-performatives-can be reg-
ulated without offending the first amendment even if speech with
similar perlocutionary effects but with different illocutionary
force cannot be so regulated.
Altman, however, does not demonstrate that hate speech
"subordinates"; rather, he merely asserts that it does. Typically,
speech acts that can be regulated without violating the first
amendment are those that invoke well-established conventions
through which legal and moral rights and duties can be altered-
conventions about promising, agreeing, consenting, waiving, giv-
ing (and accepting), swearing (an oath), and marrying, just to
name some of the more prominent. All of these conventions for
altering rights and duties are salutary, for it is a widely shared
belief that people are morally entitled to alter their rights and
duties in the respects these conventions facilitate.
Now for Altman's argument to work, there must be change
in legal or moral relationships that is effected by hate speech. As
Altman recognizes, when hate speech's illocutionary force is that
of an insult or a put-down, it fails to effect a change in legal or
moral rights and duties and inflicts only psychological harm,
harm that Altman finds insufficient to outweigh first amendment
concerns.s1 Specifically, hate speech must effect a change in legal
or moral relations that can be characterized as a subordinating
change.s2
There is, however, no recognized convention through which
people can be subordinated through speech acts. For one to be
legally subordinated, he must have his legal rights and duties al-
tered in a way that brings about a subordinate legal status. Hate
speech does not effect such an alteration. Nor does it alter moral
rights and duties. It leaves the target's moral status as it was. Of
course, those who engage in hate speech may often believe that
the targets are moral subordinates. However, they do not
believe that uttering epithets is what makes them moral
subordinates.
Altman suggests at one point that the subordinating quality
of hate speech is not its ability to alter legal or moral relations

50. Id. at 310·11.

51. ld. at 310.
52. The same problem attends the argument that pornography has the illocutionary
effect of "silencing" women. See Rae Langton, Speech Acts and Unspeakable Acts, 22
Phil. & Pub. Aff. 293 (1993) (analyzing catharine MacKinnon's position on pomogra·
phy). The examples Langton gives of subordinating speech acts-such as, "Blacks are not
permitted to vote."-all involve recognized conventions for altering legal rights. See, e.g.,
id. at 302, 304.
1996] BANNING HATE SPEECH 89

but is rather its consistency with regarding its targets as moral

subordinates.s3 Thus, if one regards blacks as moral inferiors, it
makes sense to call them "niggers" and not "moral inferiors": the
latter term invites a reasoned argument in reply, implying actual
moral equality, whereas the former is consistent with the moral
inequality asserted. The problem with this argument is that it
sweeps much too broadly. There are many ways of treating peo-
ple as if one regarded them as moral inferiors-failing to listen to
their arguments, failing to give them reasons, snubbing them so-
cially, avoiding them on the street, and so on. Acts premised on
a mistaken belief in moral inferiority may be morally wrong, but
surely we need more than moral wrongness to justify prohibiting
them, particularly when speech is involved. Hate speech does
amount to treating its targets as if they were moral inferiors, but
so too does lots of speech that we constitutionally protect.
In sum, hate speech cannot in fact subordinate. And
although it can be an apt way of expressing the mistaken moral
belief that another is morally inferior, so too can lots of speech
and conduct that are constitutionally protected.

C. HATE SPEECH AS AN Acr OF DISRESPECf

If Altman fails to show that hate speech counts as a speech

act of subordination, perhaps speech act theory can be employed
against hate speech some other way. For example, one might ar-
gue that hate speech not only expresses hatred or contempt but
also counts as the act of "treating another as hateful or contempt-
ible." Moreover, hate speech, because it violates social norms
regarding civility and respectfulness, counts as an act of incivility
and disrespect. Therefore, the argument might continue, when
we ban hate speech we are not banning ideas but banning partic-
ular kinds of socially harmful acts. Or, to use Kent Greenawalt's
felicitous phrase, when we ban hate speech, just as when we ban
perjury or criminal conspiracy, we are banning "situation alter-
ing" speech acts.54
There is no question that hate speech is a performative, and
that often its illocutionary force is that of acting disrespectfully.
In other words, hate speech not only expresses the speaker's dis-
respect of his target but also counts as treating the target disre-
spectfully. Social conventions establish the ways of
demonstrating respect and disrespect, civility and incivility, just
53. No changes in legal or moral rights and duties are effected unless the speaker
has the authority to effect subordinating changes. See id. at 311.
54. Greenawalt, Speech, Crime, and the Uses of Language at 57-63 (cited in note 29).
90 CONSTITUTIONAL COMMENTARY [Vol. 13:71
as social conventions establish how speech is to count as various
other kinds of performances.
Nevertheless, it would be both unwise and inimical to a re-
gime of free expression to allow social conventions such as those
governing respect to limit public discourse. First, a speech act
exception to freedom of speech essentially has no bounds. Al-
most any unwelcome expression could count as an act of disre-
spect, and would be characterized correctly as such whenever the
social norms such a characterization implicitly relies upon in fact
do exist. Thus, if there are social norms in Biloxi to the effect
that advocating interracial dating shows disrespect to people over
sixty, advocating interracial dating in Biloxi is a showing of disre-
spect, at least if the speaker does indeed disrespect those the
norm protects. Likewise, there is nothing to prevent Gates's
statement (A) from being deemed an act of disrespect by social
norms.
Second, as I said with respect to norms of offense, the line
between proscribing disrespectful ideas and proscribing disre-
spectful acts is one that is difficult and dangerous to draw. In-
deed, whatever other ways there are of being disrespectful,
expressing disrespect-through whatever words one chooses to do
so-would appear to be a clear case of being disrespectful.
Therefore, a ban on verbal acts of disrespect translates quite nat-
urally into a ban on expressions of [Link] And such a limi-
tation on what can be expressed in public discourse is
inconsistent with the first amendment, which is not restricted to
expression that is civil or respectful.s6

55. Consider the remarks of Robert Post on this point:

If communication could be [restricted] ... because it embodies social relations
of which we disapprove, public discourse could no longer perform [its] function
[of promoting democratic self-governance]. There is no difference ~;>etween_ ex-
cluding speech from public discourse because we condemn the sOCial relatlon-
ships it embodies and excluding speech from public discourse because we
condemn the ideas by which those social relationships are embodied.
Post, 32 Wm. & Mary L. Rev. at 299-300 (cited in note 38).
56. See David A.J. Richards, Free Speech as Toleration, in W J. Waluchow, ed., Free
Expression: Essays in Law and Philosophy 31, 47 (Oarendon Press, 1994); Post, 32 Wm.
& Mary L. Rev. at 303-04, 326-28 (cited in note 38). See also Tenniniello v. Chicago, 337
U.S. 1, 4 (1949): "[Free speech] may indeed best serve its high purpose when it ... stirs
people to anger."
The contrary view-that the law does not violate the first amendme11:t when it ~orbi_ds
disrespectful expression-is held by Delgado. See Delgado, 17 Harv. Ctv. Rts.-Ctv. Ltb.
L. Rev. at 172-78 (cited in note 20).
1996] BANNING HATE SPEECH 91

IV. THE HATE SPEECH CONTROVERSY AND FIRST

AMENDMENT THEORY: OF STICKS AND
STONES AND SLIPPERY SLOPES
Thus far I have attempted to show that although hate speech
qua epithets can cause harm by conveying painful or distracting
knowledge, hate speech is not in any way unique in its ability to
do so. Moreover, although epithets may flout conventions of of-
fense and respect, free speech cannot be subordinated to such
potentially farreaching conventions without jeopardizing its sta-
tus as an important liberty.
My basic thesis is that reflected in the childhood
countertaunt, "Sticks and stones may break my bones, but words
will never hurt me." Although most first amendment scholars
appear to take issue with this,s? in reality all they wish to point
out is that the ideas conveyed by words can be harmful, not that
the words themselves-the symbols that are the wrappers in
which the ideas are conveyed-are harmful. Yet once we focus
on this distinction between word as idea and word as symbol,
however banal it appears, we can see that attempts to expunge
harmful ideas by expunging words are both perniciousss and
doomed to failure.
In addition to attempting to locate precisely the harm in
hate speech, I have been employing a particular form of first
amendment argumentation. I have taken as a premise that gov-
ernment may not in general ban speech because the speech ex-
presses ideas that are harmful in ways X, Y, and Z. I have then
argued that the various reasons offered for banning hate speech
represent attempts to ban it because it expresses ideas that are
harmful in ways X, Y, and Z.
For someone who wishes to challenge my argument, three
possible strategies are available. First, one could argue that cur-
rent first amendment jurisprudence, with its core injunction of
content neutrality and its limited exceptions thereto, is wrong-
headed. Perhaps freedom of speech does or should rest on val-
ues that would dictate a quite different jurisprudence, perhaps
one much more sensitive to the content of speech.
I do not believe that a more content-sensitive free speech
jurisprudence is very plausible. The very point of free speech

57. See generally Frederick Schauer, The Phenomenology of Speech and Harm, 103
Ethics 635 (1993). See also Frederick Schauer, Uncoupling Free Speech, 92 Colum. L.
Rev. 1321, 1321 (1992).
58. See George Orwell, Nineteen Eighty-Four (Harcourt Brace Jovanovich, Inc.,
1949).
 92 CONSTITUTIONAL COMMENTARY [Vol. 13:71

seems to be some sort of content neutrality. And to me the most

plausible revisions of first amendment jurisprudence appear to
be those permitting fewer, not more, content-based distinctions
than does current first amendment doctrine.s9 In any event, my
interest in this Article lies not with this more ambitious revision-
ist strategy, but with two other more modest ones.
One such strategy would be to argue for a specific exception
(from first amendment protection) for epithets-the specific
words themselves. The other would be to argue for a specific
exception for the ideas that epithets express, whether or not ex-
pressed by epithets or by other words or symbols. The former is
not particularly threatening to first amendment values, but is
both unprincipled and likely to have little if any positive value.
The latter could achieve some positive results, but at the cost of
seriously undermining freedom of speech.
A first amendment exception for epithets can be dealt with
rather quickly. What would be lost in terms of positive value if
words such as "nigger," "kike," "spic," "cunt," and "faggot" were
outlawed and if the first amendment were read to permit such a
ban? Surely, not much. Anything worthwhile saying could still
be expressed in a variety of ways. The first amendment has sur-
vived the carving out of some exceptions and could survive an
additional one like this.
The real question, however, is, what would be gained? If the
same ideas that these epithets convey can be conveyed as pain-
fully, offensively, and disrespectfully in other wrappers, what is
the point of banning only epithets? Perhaps some will see a gain
in making it more difficult for the less clever and creative to spew
their venom, but I doubt that many will applaud such a discrimi-
natory free speech exception. High brow bigotry should have no
legal advantage over low brow bigotry.60

59. See, e.g., Larry A. Alexander, Trouble on Track Two: Incidental Regulations of
Speech and Free Speech Theory, 44 Hastings L.J. 921, 955-57 (1993); David A. Strauss,
Persuasion, Autonomy, and Freedom of Expression, 91 Colum. L. Rev. 334 (1991); Law-
rence Alexander and Paul Horton, The Impossibility of a Free Speech Principle, 78 Nw. U.
L. Rev. 1319, 1350-52, 1356-57 (1983); Thomas Scanlon, A Theory of Freedom of Expres-
sion, 1 Phil. & Pub. Aff. 204 (1972). See also Thomas Nagel, Personal Rights and Public
Space, 24 Phil. & Pub. Aff. 83 (1995).
60. See Gates, Speaking of Race at 47 (cited in note 16}:
If you really want to penalize . . . wounding words, it makes no se~se to
single out gutter epithets-which, on many college campuses, to be candtd, are
more likely to stipnatize the speaker than their intended victim-and leave. the
far more painful L"polite"] disquisition alone .... A .rul.e ~f thumb: in ~mencan
society today, the real power commanded by the ractst ts [Link] to .vary mve~ly
with the vulgarity with which it is expressed. . . . [T]he soctally dtsenfranchtsed
... are more likely to hail [blacks] ... as "niggers."
1996] BANNING HATE SPEECH 93

Moreover, if a list of specific words is banned, undoubtedly

new words will begin functioning as epithets. The law would
either have to be revised continually to keep up with changing
word meanings,61 or else a ban on "epithets" will have to be sub-
stituted for the ban on specific words; and a ban on "epithets"
will be both exceedingly vague and potentially quite broad.62
More importantly, a ban on "epithets" will shade naturally into a
ban on the ideas of subhuman, contemptible status that epithets
express.
That brings me to the final strategy, an exception from first
amendment protection, not for specific epithets or epithets gen-
erally, but for the ideas expressed through epithets. Why not
read the first amendment to permit government to place the
ideas that some groups of humans are moral inferiors and unwor-
thy off the table of permitted public discourse? If the expression
of these ideas, through epithets or by other means, causes psy-
chologically painful knowledge, offense, and humiliation-and if
these ideas have no countervailing positive value-why should
they be protected by the first amendment?63
Ultimately, it is this question that lies at the core of the de-
bate over hate speech. Epithets-the words themselves, the
symbolic wrappers in which ideas are conveyed-are not really
the issue. Rather, the issue is the ideas themselves. If the ideas
are wrongheaded and dangerous, what possible good can come
of granting them the exalted status of being constitutionally pro-
tected? Do they not demean the first amendment itself by claim-
ing its protection?
This attack on first amendment protection of hate speech
reveals a general difficulty with arguing for first amendment pro-
tection for any speech, particularly if the speech expresses ideas
thought to rest on misconceptions of fact or value and to be po-
tentially harmful. (The free speech defense that assumes the
ideas reflect correct conceptions of fact and value and/or are so-
cially valuable presumably has lost out in the proper forum for
the assessment of truth and value, the legislature.) What justifies
constitutional insulation of such ideas from legislative control?
61. See Richard Abel, Speech and Respect 98-107 (Stevens & Sons, Ltd., 1994) (giv-
ing several examples of how speech but not its message is altered to evade regulation).
See also Jeffrey M. Shaman, The Theory of Low-Value Speech, 48 S.M.U. L. Rev. 297,
325-26 (1995) (showing how the meaning of epithets varies with the context); Strossen,
1990 Duke L.J. at 538-39 (cited in note 27) (same).
62. See Gates, Speaking of Race at 32-33 (cited in note 16).
63. This is essentially the question Delgado raises. See Delgado, 17 Harv. Civ. Rts.-
Civ. Lib. L. Rev. at 140 (cited in note 20) (arguing that racial slurs are harmful and serve
no societal purpose).
94 CONSTITUTIONAL COMMENTARY [Vol. 13:71

One tack for defending free speech in such instances is to

trot out Mill's four arguments for protecting the expression of
falsehoods.64 Another is to cite to some strong deontological
principle, such as Scanlon's Millian principle, which denies gov-
ernment the authority to interdict expression out of fear of how it
will affect the audience's beliefs and actions,6s or Thomson's
principle denying that there can be rights against belief-mediated
harms.66
If one is not convinced by Mill, Scanlon, or Thomson, or
finds their arguments generally convincing but inapplicable to
the case at hand-here, the ideas expressed by hate speech67-
the most familiar form of free speech defense is the slippery
slope argument.68 That argument goes like this: If we permit, as
a first amendment exception, bans on hate speech, our doing so
can plausibly be read as our adopting the maxim that there is no
first amendment protection for speech expressing ideas that
cause harm and rest on incorrect facts or values. Because almost
all speech that government bans is believed by government to
cause some harm and to rest on incorrect facts or values,69 a gen-
eral first amendment exception of this type would eviscerate if
not completely swallow the first amendment. A first amendment
that protects only what is in government's eyes the true and valu-
able or the harmless is no constitutional protection worth having.
Well, then, it might be asked, if protecting the untrue, the
valueless, and the harmful is what is central to the first amend-
ment, what is so valuable about the first amendment? What do
we lose if we lose it?
Assuming the first amendment skeptic is unconvinced by
Mill, Scanlon, or Thomson, the response is that we have no rea-
son to trust in the government's competence authoritatively to
64. See John Stuart Mill, On Liberty Ch. 2 at 19, 21-67 (Bobbs-Merrill Co., 1956).
65. See Scanlon, 1 Phil. & Pub. Aff. 204 (cited in note 59). See also Nagel, 24 Phil.
& Pub. Aff. at 96-99 (cited in note 59); Strauss, 91 Colum. L. Rev. at 355-57 (cited in note
59).
66. See Thomson, The Realm of Rights at 253-59 (cited in note 41).
67. I believe that bans of epithets do run afoul of the deontological principles of
Scanlon, Thomson, and others (see the autorities cited in note 59), as well as the conse-
quentialist principles of Mill, for ultimately what offends and is disrespectful is the belief
in inferiority that lies behind the expression of the epithet, and what harms the target are
either the belief-mediated harms of offense and indignation or the tangible harm caused
by others acting on their belief in the target's inferiority.
68. See, e.g., Frederick Schauer, Slippery Slopes, 99 Harv. L. Rev. 361 (1985). For a
discussion of a related concept, see Frederick Schauer, Exceptions, 58 U. Chi. L. Rev. 871,
880-91 (1991). See also Strossen, 1990 Duke L.J. at 537 (cited in note 27).
69. Sometimes government bans speech, not because it believes the speech to rest
on incorrect facts or values, but because the speech invades privacy or infringes a prop-
erty right (i.e., copyright).
1996] BANNING HATE SPEECH 95

distinguish the true from the false and the valuable from the val-
ueless in the realm of ideas. This distrust of government-which
applies to all governmental officials, including the courts- can
never be fully vindicated, even by citing to the multitude of par-
ticular instances when government has banned the true and use-
ful under the guise of banning the false and dangerous. The
future is always a new day, and it may bring us more intelligent
and honest governmental officials than those who have come
before. And the free marketplace of ideas cannot, despite the
rhetoric to the contrary, guarantee that the true and valuable will
triumph, no matter how long the long run. Nevertheless, the first
amendment of content-neutrality-which is the first amendment
at its core-rests, if not upon broad deontological principles,
then upon the hunch that legislatures and courts should not be
trusted to distinguish true and valuable ideas from false and per-
nicious ones.7o
The skeptic may reply that although government should not
be trusted to make content distinctions generally, it can be
trusted in the case of the ideas expressed by hate speech. We
know that races, ethnic groups, religious groups, and genders are
not of unequal moral worth and that homosexuals are not subhu-
man creatures.11 Indeed, we have enshrined many of these judg-
ments of equal moral worth in the Constitution itself. Surely
here is a case where government officials have got it right when
they conclude nothing valuable will be lost if hate speech is
banned.
To strengthen this argument, let me grant not only that the
idea of moral inferiority expressed by hate speech is wrong and
thus valueless in that respect, but also that expressing that idea is
offensive and dangerous. It is dangerous because hate speech
might persuade others and lead them to act on false beliefs, to
the detriment of all concerned. Ultimately, racism and sexism
and the other similar views that hate speech expresses can never
serve as the predicates of legitimate governmental action. And
because epithets convey racist ideas in their most offensive pack-
aging, surely the case for excepting them from first amendment
protection is compelling. n

70. See Delgado, 17 Harv. Civ. Rts.-Civ. Lib. L. Rev. at 51 (cited in note 20);
Browne, 52 Ohio St. L.J. at 550 (cited in note 4); Gates, Speaking of Race at 37 (cited in
note 16); Strossen, 1990 Duke L.J. at 539 (cited in note 27).
71. See Larry Alexander, What Makes Wrongful Discrimination Wrong? Biases,
Preferences, Stereotypes, and Proxies, 141 U. Pa. L. Rev. 149, 158-63 (1992).
72. Such an argument as is set fonh in this paragraph is forcefully pressed by Alon
Hare!, Bigotry, Pornography, and the First Amendment: A Theory of Unprotected Speech,
 96 CONSTITUTIONAL COMMENTARY [Vol. 13:71
This argument, however, puts us right back on the slippery
slope. We know the ideas expressed by hate speech are vile and
potentially dangerous, you know it, and so do would-be govern-
mental banners. Yet there will be many other we's, many other
you 's, and many other officials, all just as certain about some
other speech as we are now about hate speech, and many of
these other we 's, you 's, and officials will be tragically mistaken.
The operative hunch is that it is more prudent to disable our offi-
cials even when we are certain they are right than to license them
to back up their certitude with bans on speech. It is in the end a
gamble based on nothing more than this hunch, but that may be
all that supports the first amendment and all the support it
requires.73

V. A CONSTITUTIONAL CONCESSION TO THE

PROPONENTS OF HATE SPEECH CODES
Thus far the tenor of my arguments has been one of opposi-
tion to hate speech prohibitions. I have argued that not words
but ideas cause psychological pain and sometimes concrete harm,
that therefore proscriptions of words are misdirected, but that
suppression of painful and even harmful ideas usually runs afoul
of the First Amendment.
Nonetheless, I must disappoint thoroughgoing opponents of
hate speech codes in one crucial respect. According to orthodox
constitutional doctrine, the government does not violate the First
Amendment by running schools and universities. That is, the
government does not violate the First Amendment by establish-
ing curricula and requiring its employees-public school teachers
and public university professors-to adhere to those officially es-
tablished curricula.74 This First Amendment exception to the
general principle against official orthodoxies7s is highly problem-

65 S. Cal. L. Rev. 1887 (1992). For a response to Hare!, see Alexander, 44 Hastings LJ. at
950-53 (cited in note 59).
73. Perhaps the best discussion of the general argument against entrusting govern-
mental officials with the power to ban speech based on its content is found in the works of
Fred Schauer. Schauer asks why we should be more concerned with governmental
pathologies than with speaker pathologies, and why we should be more concerned with
governmental pathologies in the regulation of speech than with governmental pathologies
in the regulation of other conduct. See, e.g., Frederick Schauer, Free Speech in a World of
Private Power, in Tom Campbell and Wajciech Sadurski, eds., Freedom of Communication
1, 10-13 (Dartmouth Publishing Co., 1994). His answer, albeit tentative, is similar to
mine: We do so based on a hunch born of our particular historical and political context.
ld. at 13-15.
74. See, e.g., Board of Educ. v. Pico, 457 U.S. 853 (1982).
75. See West Virginia Board of Educ. v. Barnette, 319 U.S. 624, 642 (1943).
1996] BANNING HATE SPEECH 97
atic in theory76 but nonetheless quite firmly entrenched in our
jurisprudence, despite the difficulties it causes in adjudicating its
margins.77
If the government may control the ideas presented in its
classrooms, then it may ban epithets in its classrooms. The harm
from epithets is still harm that comes from the ideas the epithets
express, most of which the teacher may express in other fora with
constitutionally-mandated impunity.7s (There are some special
harms that accompany classroom expression of certain ideas, pri-
marily those associated with the humiliation of being referred to
disparagingly in front of others or with the fear of unfair evalua-
tion.) Regardless whether the ideas the government wishes to
regulate in its classrooms and other teacher-student exchanges
are or are not specially harmful in those contexts, the govern-
ment may apparently regulate them to the full extent of its con-
trol over its own schools and universities.79 Therefore, one hotly
controverted arena of hate speech proscription-teacher to stu-
dent speech in public schools and public universities-is not
touched by the principal arguments in this Article invoking the
Constitution.

VI. A MORE CYNICAL VIEW OF THE URGE TO BAN

HATE SPEECH
I have acknowledged that hate speech can be harmful be-
yond its influence on third party audiences, an influence the first
amendment clearly protects. It is harmful when it conveys infor-
mation about attitudes and beliefs that is painful and not previ-
ously known, or is known but was previously subconscious and
not distracting, or known but not through a means that required
76. See S. Arons, The Separation of School and State: Pierce Reconsidered (Institute
for Humane Studies, 1977); Alexander, 44 Hastings LJ. at 953-54 (cited in note 59); Larry
Alexander, Understanding Constitutional Rights in a World of Optional Baselines, 26 San
Diego L. Rev. 175 (1989); Stanley Ingber, Socialization, Indoctrination, and the 'Pall of
Orthodoxy': Value Training in the Public Schools, 1987 U. Ill. L. Rev. 15.
77. See, e.g., Rust v. Sullivan, 111 S. Ct. 1759 (1991) (upholding content restrictions
on government funded speech); Board of Educ. v. Pico, supra note 74 (striking down
content-based removal of books from public school libraries but approving a large mea-
sure of content-based control of curriculum and of library book acquisitions); Southeast-
em Promotions, Ltd. v. Conrad, 420 U.S. 546 (1975) (striking down content-based refusal
to allow play in publicly-owned theater). See generally, Alexander, 26 San Diego L. Rev.
(cited in note 76).
78. See, e.g., Pickering v. Board of Educ., 391 U.S. 563 (1968); Larry A. Alexander,
44 Hastings L.J. at 953-54 n.113 (cited in note 59).
79. Limits were alluded to in Pico, supra note 74, but not fteshed out. See generally
Mark G. Yudof, When Government Speaks: Politics, Law, and Government Expression in
America (U. of California Press, 1983); Ingber, 1987 U. Ill. L. Rev. (cited in note 76); Cass
R. Sunstein, Government Control of Information, 74 Cal. L. Rev. 889 (1986).
98 CONSTITUTIONAL COMMENTARY [Vol. 13:71

a distracting response. But hate speech is surely not unique in its

ability to inflict these harms. Lots of speech can do so.
Although hate speech can be harmful in the ways described,
it would be unwarranted to conclude that such harmfulness is the
primary or even an important motivation for the efforts to ban it.
Indeed, it is quite possible that while hate speech can be harmful,
its actual harmfulness has been greatly exaggerated by those who
wish to ban it. In fact, I do not think that the bans on hate
speech are motivated by the desire to remain in ignorance of
others' attitudes and beliefs. Nor are those who advocate such
bans typically people who are easily and greatly offended by the
flouting of social norms, or who place a high premium on dis-
plays of civility and respect.
Nor is it likely that the principal impetus behind the moves
to ban hate speech is instead a desire to interdict possible persua-
sion of others. No doubt some do have such a concern, but hate
speech is rarely an effective means of persuading bystanders to
hold similar attitudes and beliefs.
What I would suggest regarding the movement to ban hate
speech is even more ominous. I believe it is motivated primarily
by hatred of those with bigoted attitudes and a desire to exercise
power over them. If the targets of hate speech succeed in getting
it banned, that demonstrates their own power relative to the
speakers. If they cannot get it banned, then they feel the frustra-
tion of impotence vis-a-vis those whose views they [Link]
The Skokie case is perhaps a good example. All right-think-
ing people have good reason to feel hatred toward those with
Nazi views, but perhaps none in this country have more reason to

80. See also Gates, New Republic (Sept 20 and 27, 1993) at 46-47 (cited in note 4)
(discussing the views of Richard Delgado and Marl Matsuda found in Matsuda, Words
thai Wound (cited in note 4)):
[I)n the Republic of Self-Esteem, we are invited to conceive of the lawsuit as
therapy. "When victimized by racist language," Delgado explains, "victims must
be able to threaten and institute legal action, thereby relieving the sense of help-
lessness that leads to psychological harm."
A similar therapeutic function could be played by criminal proceedings, in
Matsuda's view. When the government does nothing about racist speech, she
argues, it actually causes a secon~ injury. "The second injury is the p~in of
knowing that the government proVJdes no remedy and offers no recogmtton of
the dehumanizing experience that victims of hate propaganda are subjected to."
In fact, "The government's denial of personhood through its denial of legal re-
course may even be more painful than the initial act of hatred." Of course, what
this grievance presupposes is that the state is there, in lf!CO parentis: to confer
personhood in the first place. Finally Mats~da mu_st repatr not t? an mst~men­
tal conception of the state, but to a conception of tt as the "offietal embodtme~t
of the society we live in," which is rather. re~ote and. ai?Stracted from th~ re~h­
ties of our heterogeneous populace, wtth tts con1hctmg norms and JOSthng
values.
1996] BANNING HATE SPEECH 99

feel this way than the Holocaust survivors, their relatives, and the
other Jews who make up a large percentage of Skokie's popula-
tion. But it is doubtful that, at least after the Nazis announced
their intention to march in Skokie, many of these people were
ignorant of the fact that there were still Nazis and that those Na-
zis felt sufficient hatred of Jews to want to parade through Skokie
in Nazi regalia. Nor were many residents of Skokie particularly
concerned that a Nazi parade would distract them from other
tasks. Indeed, the effort to block the Nazis' march was probably
much more distracting. What seems much more likely is that
stopping the march would frustrate the hated Nazis, who could
not otherwise be punished just for being Nazis, and that allowing
them to march would frustrate the quite natural urge to strike
out at those whose ideas one hates. In other words, hate speech
situations like Skokie are frequently perceived by the targets of
hate speech as zero sum games in which the only question is
which of the antagonistic groups will prevail. If the hate speech
is allowed, the speakers will have prevailed. Therefore, the
targets, to prevail, must suppress the speech.s1
I admit that I am engaged in the worst kind of armchair psy-
choanalysis, and that I have nothing more solid to back me up
than hunches.82 Nevertheless, if my hypothesis is correct, then

81. Compelling a showing of "respect" from those who do not in fact respect one is
a way of punishing the latter for their beliefs and perhaps for eliciting respect for one's
power if not one's humanity.
82. Those hunches, however, are shared by others. Gates notes:
In fact, the main appeal of speech codes usually turns out to be primarily expres-
sive or symbolic rather than consequential in nature. That is, their advocates do
not depend on the claim that the statute will spare victim groups some foresee-
able amount of psychic trauma. They say, rather, that by adopting such a stat-
ute, the university expresses its opposition to hate speech and bigotry. More
positively, the statute symbolizes our commitment to tolerance, to the creation
of an educational environment where mutual colloquy and comity are
preserved ....
And yet once we have admitted that the regulation of racist speech is, in
part or whole, a symbolic act, we must register the force of the other symbolic
considerations that may come into play. So, even if you think that the notion of
free speech contains logical inconsistencies, you need to register the symbolic
force of its further abridgement. And it is this level of scrutiny that may tip the
balance in the other direction. The controversy over ftag burning is a good illus-
tration of the two-edged nature of symbolic arguments. Perhaps safeguarding
the ftag symbolized something nice, but for many of us, safeguarding our free-
dom to burn the ftag symbolized something nicer.
Note, too, the contradiction in the expressivist position I just reviewed: a
university administration that merely condemns hate speech, without mobilizing
punitive sanctions, is held to have done little, offering "mere words." Yet this
skepticism about the potency of "mere words" comports oddly with the attempt
to regulat~ "mere words" that, since they are spoken by those not in a position
of authonty, would seem to have even less symbolic force. Why is it "mere
words" when a university condemns racist speech, but not when the student ut-
 100 CONSTITUTIONAL COMMENTARY [Vol. 13:71

banning hate speech will be only a small, appetite-whetting vic-

tory for forces of censorship. For as I have pointed out, lots of
speech beyond hate speech reveals hateful attitudes and identi-
fies those whom others will now hate in return.
All of this is quite human and natural, if regrettable. The
First Amendment is in the Constitution to check what is quite
natural, particularly the urge to punish those with whom you dis-
agree because you hate or fear those whose thoughts you find
hateful or dangerous. Knowledge of hateful attitudes and beliefs
can be painful and thus harmful in that sense, but I think it is
unlikely that this pain explains the present movement to censor.
The desire to punish, embarrass, and bend to one's will those
who hate you is the most plausible explanation.83 Aversive
thought control and retaliation are, however, unlikely candidates
for the status of legitimate reasons for suppressing speech.84

ters the abusive words in the first place? Whose words are "only words"? Why
are racist words deeds, but anti-racist words just lip service?
Further, is the verbal situation as asymmetric as it first appears? Does the
rebuke "racist" have no power to wound on a college campus?
Gates, Speaking of Race at 38-39 (cited in note 16).
83. See Abel, Speech and Respect at 22-29 (cited in note 61); James B. Jacobs, The
Emergence and Implications of American Hate Crime Jurisprudence, 22 Israel Yearbook
on Human Rights, 113, 136 (1993).
84. But see Delgado, 17 Harv. Civ. Rts.-Civ. Lib. L. Rev. at 149 (cited in note 20)
(arguing that if the government can change expression it can change thoughts, and that
this is a justification for suppressing speech).