In my Opening Article I explored the potential of “upstream” speech restrictions to undermine the political legitimacy of “downstream” laws. Using hate speech bans as an example, I argued that these restrictions had the potential to seriously compromise, and in some cases even annihilate, the legitimacy of antidiscrimination laws as applied to those whose ability to publicly object to these downstream laws had been impaired by the speech bans. Professor Jeremy Waldron, whose criticism of Professor Ronald Dworkin’s and my previous and rather cursory presentations of this proposition inspired me to more fully explore and develop this position, wrote a response to my Opening Article. A group of distinguished scholars from several nations and working in various disciplines then commented on this discussion.
Admirably fulfilling his role as my principal interlocutor, Waldron filed a vigorous and comprehensive conceptual challenge to the position that upstream speech restrictions can deprive downstream laws of legitimacy. Although none of the commentators fully embraced Waldron’s wholesale rejection of my position, several offered trenchant challenges to some specific aspects of my argument. But the brunt of the disagreement with my Opening Article was empirical and related specifically to hate speech. Several commentators thought that I exaggerated the extent of the restriction that hate speech bans had placed on the ability of people to oppose antidiscrimination measures. They also claimed that I underestimated the effect on legitimacy resulting from hate speech itself. There was also vigorous disagreement about the appropriate legal response in a state of uncertainty about the harmful effects of hate speech.
In this Reply I will address both the general criticism of my view that viewpoint-based restrictions on public discourse can deprive downstream laws of legitimacy, as well as the specific criticism of my claims about the detrimental effect on the legitimacy of antidiscrimination laws resulting from hate speech bans. But before doing so, I think it might be helpful to explain why I chose hate speech as the exemplar of speech restrictions with the potential to undermine the legitimacy of downstream laws. I will also address a concern raised by Professor Frederick Schauer about this choice.
To explore the idea that upstream speech restrictions can impair the legitimacy of downstream laws, I focused on hate speech bans for several reasons. First, I was aware of no other type of speech restriction commonly applied (or misapplied) in liberal democracies as likely to have this effect. In addition, as just mentioned, the first person to critically engage the proposition that a speech restriction could deprive a downstream law of legitimacy was Waldron, who did so in two works defending narrow restrictions on hate speech. Finally, as Professor Vincent Blasi appreciates in his Commentary, I was concerned that hate speech bans enacted to protect members of vulnerable minorities may have the perverse effect of impairing the legitimacy of antidiscrimination measures enacted to protect these same individuals.
In his Commentary, Professor Schauer expresses the concern that the focus on hate speech restrictions risks that “the analysis of interesting and important questions about the relationship between political legitimacy and freedom of speech will be both crowded out and distorted” by this contentious subject. Schauer’s fear has, at least to some extent, materialized: the discussion has morphed into one in which the propriety of hate speech bans in liberal democratic societies has assumed equal billing with the relationship between free speech and political legitimacy. This shift in emphasis has undoubtedly, as Schauer predicted, resulted in a somewhat less sharp focus on the relationship between free speech and legitimacy than I had hoped for. On the other hand, this development may at the same time have been beneficial if, as two commentators have claimed, examining hate speech regulation through the lens of political legitimacy has moved the stale and stalled discussion of propriety of hate speech in a novel and helpful direction. In Part I of this Reply I will try to mitigate the problem that Schauer identified by addressing with as few references to hate speech as possible Waldron’s largely conceptual objection to my argument. I will also reply to some more specific objections to the view that speech restrictions can rob downstream laws of legitimacy. I will then in Part II reply to various criticisms of my argument that hate speech restrictions as they actually operate in many democratic societies have undermined the legitimacy of antidiscrimination measures.
- Challenges to the Claim that Upstream Speech Restrictions Can Deprive Downstream Legislation of Legitimacy
- Waldron’s Critique of the Claim that Upstream Speech Restrictions Can Render Immoral Enforcement of Downstream Legislation
As already noted, Waldron is the only participant in this Symposium to file a wholesale challenge to the position that speech restrictions can impair, and in some instances even destroy, the legitimacy of a downstream legislation as applied to citizens whose ability to speak out against the legislation was impaired by the speech restriction. It should be noted, however, that Waldron challenges only the normative aspect of my claim. He does not take issue with the contention that upstream speech restriction can have a detrimental effect on the descriptive legitimacy of downstream legislation.
The idea that I developed and defended in my Opening Article concerning the downstream effect of upstream speech restrictions was a novel one, apparently more so than I appreciated, and as such, required rigorous testing. I am therefore most grateful to Waldron for providing a thoughtful, comprehensive, and vigorous critique of this position. As a preliminary matter, though, it is worth noting that Waldron may characterize my position somewhat more strongly than I meant to express it.
Waldron writes that I contend that an upstream speech restriction, Lu, can so severely impair a particular person’s, P’s, ability to speak out against a proposed downstream law, Ld, that P has “a right that it not be enforced against him.” But this paraphrase is not quite accurate, for I said nothing about the P having a “right” not to have Ld enforcement against him; rather, I claimed only that such enforcement was immoral. I do not believe it to be the case that people necessarily have even a moral right not to have immoral laws enforced against them. Still, given the nature of the deprivation described in my Evangelical Photographer Scenario, it is a fair inference from what I wrote that Elaine does have a moral right not to have the antidiscrimination law enforced against her for refusing to photograph a same-sex marriage. Where Waldron may overstate my position is if he reads me as contending that this moral right necessarily implies that there should be a corresponding positive law right against enforcement. While I think there should be a presumption in any legal system against the enforcement of laws that violate moral rights, I do not believe that in every such instance people should necessarily have a positive law right against such enforcement. This is because in my view doctrinal “fit” as well as morality should play a role in determining whether a right should be recognized.
In addition, and more significantly, Waldron’s imputation of a rights generating effect of my position talks past an important aspect of the claim. Although I believe that under certain circumstances a viewpoint-discriminatory speech restriction can indeed render immoral enforcement of a downstream law, much more commonly such restrictions will diminish but not utterly destroy the legitimacy of the Ld’s enforcement. So to the extent that Waldron’s criticism of my position derives from what he sees as its improbable “deontic” consequences, it fails to engage my claim that upstream speech restrictions can impair normative legitimacy short of destroying it. For such non-lethal damage to legitimacy, while something to “regret,” has no particular implication for generations of rights, moral or positive.
With these clarifications of my position, we are now in a position to evaluate Waldron’s critique of it.
There are several crucial flaws in Waldron’s analysis that undercut his conclusion that upstream speech restrictions cannot render immoral the enforcement of downstream legislation against particular individuals. Waldron begins his critique by asserting that “debates and decisions in a representative legislature are usually seen as legitimizing the enforcement as law of the bills that survive this process.” For a moment, then, it seems as if Waldron is going to assert that the legislative process is alone sufficient to legitimize any law resulting from this process, even one with which people can reasonably disagree. And perhaps Waldron in fact holds this view. He decides, however, not to rest his entire case on such a parsimonious and highly contestable view of the relationship between free speech and political legitimacy. Rather, Waldron allows that the “best case” for my argument “looks at the informal public debate that is involved in the election and electoral accountability of legislators and in the debates in the community that complement legislative debates in the parliament.”
But even on the assumption that public discourse is “an indispensable part of the political process,” the restriction of which might perhaps have some impact on downstream legislation, it is “quite another thing” in Waldron’s view to assert that legal restrictions on this “chaotic and unformed” public debate can have the direct consequences for legitimacy that I claim. For even were a speech restriction to have a “deleterious impact on the quality of the political process,” the difficulty lies in “[i]ndividualizing its moral effects to generate particular rights” of an individual to disobey the law or to not have it enforced against him.
To try to show that there is not “any good political argument” for this position, Waldron considers the effect on legitimacy of someone being wrongfully disenfranchised. He references laws in the United States restricting early voting and imposing onerous voter ID-requirements. Assuming that these restrictions unjustifiably prevented people, including a particular person “Q,” from voting, this in Waldron’s view may be a “deplorable state of affairs.” Still, Waldron continues,
few people believe that any of the laws enacted by the legislature (to whose membership Q’s vote might have made the sort of difference that individual votes make in elections) are rendered illegitimate . . . so far as Q is concerned. No one thinks Q now has the right to disobey the laws or not have them enforced against him. His disenfranchisement may make the democracy poorer, and Q certainly has a justified complaint; but nothing follows about legitimacy and enforcement so far as his relation to the laws is concerned.
Waldron asserts that if nothing follows about the morality of enforcement of the laws against Q in the “relatively formalized context of voting,” then it cannot “possibly be true” that restriction on P’s participation in the “diffuse free-wheeling debate” that characterizes public discourse could create any legitimacy problem in enforcing a downstream law against P. As I shall discuss in detail in a moment, Waldron’s disenfranchisement scenario is in a crucial respect disanalogous to the circumstances under which I claim that a speech restriction can render immoral the enforcement of a downstream law. But first I think it would be useful to bring to the surface and then contest a crucial assumption underlying Waldron’s argument.
Waldron assumes that voting is categorically more essential to political legitimacy than is participation in public discourse. Such a hierarchy is implicit in Waldron’s assertion that if the disenfranchisement of Q does not raise legitimacy concerns, then a restriction on P’s participation in public discourse could “not possibly” do so. Waldron does not purport to make a sustained argument in support of this view. Instead, he emphasizes that although an individual’s contribution to “the swirling maelstrom of informal debate” might have some effect on “things that are said and votes that are cast in the legislature,” there is no guarantee that it will have any effect whatsoever. This is because, as Waldron explains: “My letter to the newspaper may not be published; there may be no hits on my blog; eyes may be turned away from my graffiti; my spoken words may disappear into the wind; perhaps no one will turn up for the meetings I organize; and the leaflets I distribute may end up in the gutter . . . .”
It is true that an individual’s contribution to public discourse will often have no effect on the legislative process, on who is elected to the legislature, or in persuading other individuals about matters of public concern. However, it is also true that an individual vote will almost never make a difference to the outcome of the election. Moreover, while most contributions to public discourse will when viewed in isolation have no effect on society’s collective decisions, and few will have more than negligible effect, sometimes an individual contribution, such as a blog post that goes viral or a particularly persuasive newspaper column, can have a marked effect on the outcome of such decisions. In contrast, a single vote can never make such a disproportionate contribution to the result of a collective decision. So even if the key measure of the legitimating function were, as Waldron suggests, the effect on collective decision making, it is not at all clear that public discourse ranks lower than voting.
But more significantly, even if it were the case that voting was categorically more important to legitimation than participation in public discourse, Waldron’s disenfranchisement scenario cannot do the work he intends for it. The type of voting restriction Waldron references is simply not analogous to the viewpoint-discriminatory laws that I argue can render the enforcement of a downstream law immoral as applied to particular individuals. While Waldron stipulates that the voting restrictions in his scenario are unjustified, he does not posit that they were intended to disenfranchise Q or anyone else on account of ideology or for holding a viewpoint on a particular issue.
A much more analogous scenario would be voting restrictions for an upcoming city council election cleverly designed to disenfranchise as many people as possible who oppose an increase in the property tax. Under such circumstances, I would think that Q1, disenfranchised because of her opposition to the tax increase, would have a very good claim that a law passed by the newly-elected city council raising the tax on her property was immoral. And this would be true even if the property owner could not show that but for her particular disenfranchisement the city council would have not passed the tax increase. By the same token, P1, who was able to vote in the city council election but who was forbidden by law from speaking out against the tax increase in “the maelstrom of informal debate” would, like Q1, also have a strong moral objection to the application of the tax increase to her.
Waldron correctly notes that the crux of my claim that speech restrictions can render the enforcement of downstream laws immoral is the basic precept that “‘each individual in society is of equal moral worth and therefore is entitled to have his or her interests treated with equal respect by the government.’” But it is precisely this equality concern that is missing from Waldron’s disenfranchisement scenario. From an equality perspective, it is one thing to be disenfranchised by unnecessarily burdensome restrictions on early voting or unreasonably onerous ID-requirements; it is quite another to be disenfranchised because of one’s position on a particular collective decision. In my view, there comes a point at which a speech restriction, like selective disenfranchisement, can so profoundly disrespect both the interests and equal moral worth of some individuals that the restriction can have an effect not just on the legitimacy of the legal system but also on particular laws enforced against those whose ability to oppose these laws was severely curtailed.
Waldron apparently rejects my equality argument on the grounds that even if a speech restriction were to disrespect a particular interest of the would-be speaker, there is a host of other interests “to be served by our laws” such as “health care, education, roads, housing” which the speech restriction does not implicate. That there are numerous interests not affected by the speech restrictions explains why in mature, stable democracies even the most restrictive viewpoint-discriminatory speech restriction does not come anywhere close to destroying the legitimacy of the entire legal system. But the existence of all these other legally-conferred benefits, while relevant to the measure of a legal system’s legitimacy, does little to ameliorate the damage to a citizen’s ability to protect those interests associated with a particular downstream law arising from an upstream speech restriction.
Believing that concern for the interests of the would-be speaker will not yield the detrimental effect I claim that speech restrictions can have on downstream laws, Waldron concludes that the essence of my equality concern must be about “respect for opinions.” “The main way in which we express people’s opinions in the political process,” Waldron insists, “is by counting their votes, and we do count the votes of those whose free expression is impacted” by speech restrictions. So in the end we have come full circle: it is primarily Waldron’s disagreement with my view about the legitimating power of free speech as compared to voting, not some basic conceptual flaw in my argument, that leads him to reject my view about the potentially detrimental effects that speech restrictions might have on the legitimacy of downstream legislation. Indeed, I suspect herein lies much of the reason for our differing views about the propriety of hate speech bans in a free and democratic society.
- Other Challenges to the Basic Claim
- Rule-of-Law Problems
Waldron says that he is worried that my view that an upstream speech restriction can render immoral the enforcement of downstream laws against some people but not others “may get tangled up in Rule-of-Law issues about generality.” He notes that speech restrictions are usually quite general and although they may be designed to be enforced against only certain speakers, “they have a potential impact on everyone’s speech.” As a result, he fears “it may be hard to identify the basis for in personam illegitimacy of the type” that I suggest.
I agree that my argument raises rule-of-law issues and not just the one that Waldron notes. In addition to the identification problem he worries about, there is the equal justice concern of a particular class of people possibly being exempted from obligations imposed on everyone else in society. But if I am right that upstream speech restrictions can render the application of downstream laws immoral, then criticizing my argument for any rule-of-law problems resulting from this unfortunate consequence is a classic case of “shooting the messenger.” Surely the blame for any such problems lies with the speech restrictions. In any event, these rule-of-law problems are essentially no different than the ones that commonly arise when general laws provide exemptions for a certain class people, e.g., those with religious objections to the law, or when courts find that exemptions are required by constitutional provisions, such as guaranteeing the free exercise of religion. In both instances, there is the problem of identifying who is eligible for the exemption and justifying the impact on the equal administration of the laws arising from the exemption.
In fact, any-rule-of law issues raised by my argument are less problematic than the exemptions from general law just discussed. Aside from providing a reason against the enactment of viewpoint-discriminatory laws, the only other practical consequence of my analysis is to detect instances in which the enforcement of a downstream law may have been rendered immoral by an upstream speech restriction. As I have explained, however, this does not necessarily mean that the law should not be enforced. And where the law will be enforced despite the moral deficit, there is, unlike with granting an exemption, no need to identify with precision those against whom enforcement of the law would be immoral.
- The Relationship between the Justification of a Speech Restriction and its Effect on Political Legitimacy
I concluded my Opening Article by arguing that the untoward effects that hate speech restrictions have on the legitimacy of antidiscrimination laws “weigh against such upstream constraints.” In his Commentary, Professor Steven Shiffrin succinctly offers a powerful conceptual challenge not just to this specific conclusion regarding hate speech laws but also to the significance of my entire position as it relates to normative legitimacy. “If hate speech restrictions are justifiable,” Shiffrin writes, “then their enforcement cannot undermine the normative legitimacy of anti-discrimination laws.” If, on the other hand, “such restrictions are not justifiable,” Shiffrin contends, “then the impact on anti-discrimination laws is interesting, but not central to the case against them.” For this reason, Shiffrin characterizes my argument as “an instance of the tail wagging the dog.” Waldron makes a similar argument, observing that since “it is only unjustified restrictions on speech that affect [normative] legitimacy, then it looks as though we will have to settle the question of justification first, before we assess the impact on legitimacy.” He adds that for this reason “the argument about legitimacy can hardly be cited as a reason for thinking” that an upstream law is inappropriate.
There are two problems with this interesting conceptual challenge. First, it mistakenly assumes that whether a law is justified is an all-or-nothing proposition, when as Waldron helpfully emphasized regarding the effect of speech restrictions on legitimacy, “the legitimacy of any given law is itself a matter of degree.” Relatedly, the argument ignores the iterative nature of the process for determining whether a law is justified.
To say that a law or a proposed law is justified means that there are better reasons for the law (its benefits) than there are reasons against it (its costs). So while in the final analysis a law is either justified or it is not, it is also the case that some laws are better justified than others. With respect to speech restrictions that are extremely well justified, Shiffrin and Waldron are surely correct that such constraints can have no detrimental effect on normative legitimacy. To borrow a venerable example from Justice Oliver Wendell Holmes, a law punishing falsely crying “fire!” with the intent of causing a panic would have no impact on normative legitimacy. In contrast, speech restrictions that on balance may be justified because they are necessary to prevent serious harm but nonetheless trench upon core free speech interests undoubtedly can have serious detrimental effects on descriptive legitimacy. And it could be argued that such laws impair normative legitimacy to some extent as well.
Consider, for instance, the Smith Act, a law passed early in the Cold War banning advocacy of the overthrow of the United States government by force of violence. While I do not believe this law was justified, many thoughtful people, including a majority of the Justices of the United States Supreme Court, believed that it was. But even if this law was justified, the cost to freedom of expression cannot be seriously doubted. As Justice Felix Frankfurter candidly acknowledged in voting to uphold the convictions of the leaders of the American Communist Party for conspiring to engage in advocacy forbidden under the Smith Act:
[C]oupled with such advocacy is criticism of defects in our society. . . . [Moreover, suppressing] advocates of overthrow inevitably will also silence critics who do not advocate overthrow but fear that their criticism may be so construed. . . . [It] is self-delusion to think that we can punish [the defendants] for their advocacy without adding to the risks run by loyal citizens who honestly believe in some of the reforms these defendants advance. It is a sobering fact that in sustaining the convictions before us we can hardly escape restrictions on the interchange of ideas.
So even if the Smith Act was justified, by chilling honest criticism of “defects in [American] society,” the law nonetheless had a detrimental effect on the legitimizing function of freedom of expression, certainly in the descriptive sense, and arguably in the normative sense as well. Similarly, even if the restriction on core political speech imposed a century and a half earlier by the Alien and Sedition Act was justified, as Waldron has suggested, the law damaged descriptive legitimacy and arguably negatively affected normative legitimacy. So it is not at all certain, as Waldron and Shiffrin assert, that only unjustified laws can have a detrimental effect on normative legitimacy. Rather, the better view may be that if speech restriction justification is a close call, the law can impair normative as well as descriptive legitimacy.
There is, however, a contrary view with considerable force. If under the best assessments that can be made at any particular time a law is justified, then, so are any costs, including the “chilling effect,” it may impose. On this view, if the Smith Act and Alien and Sedition Act were justified, then even the considerable cost to descriptive legitimacy, though regrettable, was not wrongful and thus did not damage normative legitimacy. But even on this view, a law’s justification is not a static, one-shot determination but is rather a dynamic and continuous process. So even if Waldron is right that we will “have to settle the question of justification first, before we assess the impact on legitimacy,” he fails to take into account that any determination about whether a law is justified, especially one made before a law goes into effect, must be open to revision in light of information about its actual operation. With respect to speech restrictions, such reevaluation would include how effective the law has been in preventing the harm thought to be caused by the restricted speech as well as whether the law has been misapplied, thereby having a greater “chilling effect” on non-targeted speech than anticipated.
In my Opening Article, I cited a number of cases in which hate speech laws were applied, or misapplied, to speech that must be allowed in any free and democratic society and thus wrongfully punished people for participating in public discourse. In addition, as a result of such misapplication, would-be speakers were likely wrongfully deterred from engaging in this public debate in ways likely not accurately predicted in the initial assessments of whether these laws were justified. In light of this new information, the justification of at least those hate speech laws that have been applied in this way, and perhaps all those of its genus, should be reevaluated in light of the detrimental effect on legitimacy.
There is another reason for reevaluation of the justification of hate speech laws and hence their potential impact on normative legitimacy. Even according to its harshest critic, the claim that upstream speech restrictions can deprive downstream laws of political legitimacy is a novel one. Accordingly, if this claim has any validity, then the justification of hate speech laws should be reassessed to evaluate the possibility of untoward effects of these laws that had not been previously considered.
Professor Eric Heinze’s Commentary also raises the relationship between justification and legitimacy, though in a very different way than Shiffrin’s and Waldron’s critiques. The burden of Heinze’s learned Commentary is to show that viewpoint-discriminatory restrictions on public discourse such as hate speech bans are illegitimate. Problematically, he repeatedly attributes this view to me, and at one point even refers to it as my “thesis.” My thesis, however, focuses not on the legitimacy of upstream laws that impose constraints on public discourse, as does Heinze’s Commentary, but rather on the effect that these constraints might have on the legitimacy of downstream legislation. In my Opening Article I contended that the potentially baleful effect that hate speech restrictions could have on downstream legislation “weighs against” the propriety, i.e., the justification, of such laws in a free and democratic society. I said nothing about the legitimacy of such speech restrictions themselves. And surely just because a law, on balance, is not adequately justified does not mean that it is also illegitimate, at least not in the sense that I have used the term. Concededly, because of the negative impact that such speech restrictions have on democracy it may well be that inadequately justified viewpoint-discriminatory laws are illlegitimate under some accepted definition of that term. This is an interesting question worthy of exploration. But with so many other issues to consider in this Symposium, it is not one I chose to investigate.
- Challenges to Specific Aspects of the Claim
On an understanding of normative legitimacy as “a fundamentally non-consequentialist and non-instrumental idea” as well as one focusing “primarily on procedure in the broadest sense of that word,” Professor Frederick Schauer agrees that my view about free speech and normative legitimacy seems “largely correct.” Specifically, on such an understanding he agrees that “allowing people to object to policies with which they disagree is a necessary component of normative legitimacy, and thus of the warrant of the state to enforce its directive by coercive means.” His one caveat is that it is “plausible” to argue that so long as citizens have the right to choose who will represent them, the “right of the citizen to speak out” is not a necessary condition of democratic legitimacy. Schauer mentions this possible objection in order to emphasize that “a strong and continuous right to freedom of speech is not entailed by the very idea of democracy, at least as long as the idea of representative democracy is not an oxymoron.” Despite this observation, however, Schauer acknowledges that it is “difficult to imagine a process of selecting representatives or policies that is not crucially facilitated by direct citizen speech” and that it is even “more difficult” to envision a government that “does not permit those people to participate in policy-making outside of the episodic process of voting.” He thus concludes that “freedom of speech is arguably simply part of the definition of what democracy and democratic legitimacy just are.”
Given the role that public opinion plays in determining the results of collective decisions in every democratic society that exists in the world today, I have no sympathy for the view that the right of the people to elect their representatives is alone sufficient to legitimize the use of coercion to enforce these decisions. It was therefore useful for someone with a bit more sympathy for this position to present it—and then largely refute it. This is especially true since my principal interlocutor seems to have a lot of sympathy for this position.
With respect to free speech and descriptive (or sociological) legitimacy, Schauer thinks it “entirely reasonable” for me to conclude based upon extrapolation from empirical studies that people claim to be more willing to comply with laws with which they disagree if they have had an opportunity to publicly express their disagreement with these measures. Still, Schauer notes that the empirical research on which I rely “stops short of answering the ultimate question—will people who claim to believe in the obligation to obey laws with which they disagree, and who in fact believe that they have an obligation to obey such laws, actually obey such laws”? Schauer correctly observes that just because people believe a course of action is proper does not mean that they will in in fact follow that belief. He thinks that this is especially true where there are “temptation[s]” to deviate from that belief. Since disagreement with a law is such a “temptation,” Schauer concludes that the “abstract belief” in obeying such laws may well be overwhelmed in “actual practice.” Because we lack empirical evidence that the belief that a law ought to be obeyed translates into compliance with the law, in Schauer’s view all we really can say about the effect of free speech on this crucial question regarding descriptive legitimacy is that “we do not know.”
I agree with Schauer that I did not attend carefully enough to the distinction between belief in complying with a disagreeable law and actually doing so. I did not mean to claim, as a statement in my Opening Article can reasonably be read as suggesting, that there was a perfect or even a strong correspondence between such belief and practice. Contrary to Schauer, however, I think we can with some confidence conclude that to some as yet unascertained extent there is, though certainly not in every case, a causal connection between such belief and actual compliance. To use Schauer’s example, a person’s belief that he should lose weight is certainly no guarantee that he will do so given temptations such as the ready availability of tasty, highly caloric food. But all things held equal, including those temptations, it would seem reasonable to conclude that the people who have a belief that they should lose weight are more likely to do so than those who have no such belief.
As I noted in my Opening Article, Schauer has in recent work helpfully clarified what it means to “obey the law” and has usefully criticized empirical studies for their confusion on this issue. I am pleased that Schauer has used his Commentary in this Symposium to further advance his important project.
Like Schauer’s commentary, Dean Robert Post’s contribution focuses on descriptive legitimacy. But unlike Schauer, he chooses to concentrate on the damage speech restrictions, including hate speech bans, can have on systemic legitimacy as opposed to the legitimacy of particular laws. In Part II, I will discuss Post’s incisive discussion of the impact of speech restrictions on systemic legitimacy. But here I want to engage the few brief comments he does make about the potential of speech restrictions to undercut the legitimacy of particular downstream laws.
Post agrees that it “is conceivable that regulations of speech might undermine the descriptive legitimacy of specific laws.” With respect to hate speech bans, for instance, he believes that if such restrictions are “capaciously interpreted” in the “distressingly broad ways” that my Opening Article shows they can be, then downstream antidiscrimination measures “may well suffer from diminished legitimacy” by making those prevented from expressing opposition to the law likely to regard the laws as “unfairly enacted.” Nonetheless, Post considers the detrimental effect on legitimacy might have on individual laws as “a complicated and largely idiosyncratic question.” He contends that “[i]ndividual laws become descriptively illegitimate primarily because they are mismatched to the mores of the population to which they are applied.” As examples, Post recalls that the Constitutional Amendment establishing Prohibition in the United States was widely regarded as illegitimate in the Northeast, while the Amendment prohibiting racial discrimination in voting was long treated as illegitimate in the South.
The idiosyncratic nature of descriptive legitimacy of individual laws that Post notes, however, has to do with profound disagreement about the substance of laws. The damage to legitimacy that can result from speech restrictions, in contrast, is a procedural concern arising from the curtailment of the ability of dissenters to express opposition to the law. Though the opposition to the substance of various laws might be idiosyncratic, there is nothing idiosyncratic about the feeling that a law has been “unfairly enacted” because one’s ability to oppose it was selectively curtailed. These feelings are no more idiosyncratic than those of “persons of widely varying views” coming to “distrust a political system that holds out the promise of self-determination but that refuses to hear what [they] have to say.”
I accept, as Post implies, that it might be “complicated” to disentangle the damage to descriptive legitimacy caused by the speech restriction from that resulting from vehement opposition to the substance of the law. But as Post correctly recognizes, unlike detriments to normative legitimacy, damage to descriptive legitimacy does not support arguments that there is no obligation to obey the law. Accordingly, the procedural and substantive elements need not be neatly sorted out to appreciate that curtailing the ability of citizens to express opposition to a law can greatly exacerbate damage to descriptive legitimacy arising from profound disagreement with the law’s substance.
- Hate Speech Bans and Political Legitimacy
In this Part, I will respond to criticisms that I have exaggerated the extent to which hate speech bans have impaired citizens’ ability to participate as equals in public discourse, while at the same time failed to appreciate the effect that hate speech itself has on political legitimacy. But before doing so it might be useful to clear up some possible misconceptions about the goals of my Opening Article as well as about my position on hate speech regulation more generally.
- The Goals of the Opening Article
First, I want to emphasize that the primary purpose of my Opening Article was to expand and explore in detail the argument, stated only cursorily in previous works, that upstream speech restriction could damage, and potentially even destroy, the legitimacy of downstream legislation. For the reasons explained above, I chose to use hate speech bans as the exemplar of speech restrictions that could have this baneful effect. Though the discussion has to some significant extent shifted in that direction, it was not my intent in my Opening Article to engage in a comprehensive discussion of the arguments for or against hate speech regulation or to argue that hate speech regulation is “wrong in principle.”
Professor Katharine Gelber complains that my Opening Article “lacks a clear conception of hate speech.” But precisely because this article was not meant to be a comprehensive discussion of the pros and cons of hate speech legislation, I did not think it necessary to undertake the difficult task of defining hate speech. In any event, the lack of “a clear conception of hate speech” in the Opening Article seems not to have been problematic. Crucially, no participant in this Symposium gainsaid that the speech in my key Evangelical Photographer Scenario might have been prosecuted in some democratic jurisdictions as hate speech.
Similarly, I hope I might be acquitted for not explaining “the defining features of hate speech, properly understood,” by which Professor Gelber evidently means expression that can and should be properly banned in a mature and stable democratic society. While I am open to the possibility that hate speech restrictions in such societies might under certain circumstances be appropriate, I remain skeptical of their propriety. So asking me to explain the defining features of hate speech that can be properly suppressed is not unlike asking an agnostic to explain the essential characteristics of God.
Having dealt with a possible important misconception of the intent of my Opening Article, I turn now to Waldron’s questioning my motive for focusing on the detrimental effects that hate speech bans can have on political legitimacy. In his Response, Waldron writes:
Back of all the points I am going to make in this essay responding to Weinstein is a worry that the argument about political legitimacy is just being wheeled into the hate speech debate opportunistically by people who have never otherwise shown that they take it seriously. I want to make sure that the argument is not just being rigged up for the purposes of the hate speech debate.
Waldron needn’t have worried. The relationship between free speech and legitimacy has long been a particular interest of mine, a connection that I have explored in works far afield from the “the hate speech debate.” For instance, in an article in another Symposium in this journal, I defended at length political legitimacy as the normative essence of American free speech doctrine. And far from being “rigged up” to specifically protect hate speech, I emphasized in these previous works that the potential for a speech restriction to damage the legitimacy of downstream legislation applies to any viewpoint-based restriction on public discourse—be it a restriction on anti-war speech, opposition to a tax increase, or criticism of a nation’s immigration policy.
Finally, I want to dispel any notion that I believe hate speech is incapable of causing harm beyond profound offense. To the contrary, as I have often explained, I believe that even in a mature, stable democracy hate speech has the potential to produce any number of much more serious harms including discrimination and even violence against those denounced by the bigoted expression. In addition, as I explained in my Opening Article, I agree with Waldron that hate speech can make vulnerable minorities unsure of their status in society. And as I also acknowledged in that article, hate speech might even damage the political legitimacy that I argue is undermined by hate speech laws, a concern I will address in detail later in this Reply.
With these possible misconceptions cleared up, I will now turn to the two main criticisms of my claim that hate speech bans impair the legitimacy of downstream legislation, particularly antidiscrimination measures.
- The Nature and Extent of Hate Speech Restrictions
Several symposiasts claim that I exaggerate the detrimental effect that hate speech bans can have on the legitimacy of downstream legislation. Some of the criticism focuses on the nature of hate speech bans and deny my contention that such bans are viewpoint discriminatory. Other critiques center on the extent of the restriction imposed by hate speech laws and contend that it is far less than I claim. I will deal with each of these points in turn.
- The Nature of the Curtailment
Central to my view about the relationship between free speech and political legitimacy is that viewpoint-discriminatory speech regulations, that is, laws that forbid people from expressing certain viewpoints in public discourse, can have grave consequences for the legitimacy of downstream legislation. In my Opening Article, I cited hate speech bans as an exemplar of viewpoint-discriminatory regulation and explored the detrimental effect these upstream speech restrictions could have on downstream legislation, particularly antidiscrimination laws.
In his Response, Waldron takes issue with my characterization of hate speech laws as viewpoint discriminatory. Indeed, qualifying his previous position that hate speech laws are “undoubtedly content-based,” he now argues that they are only “sort of content-based.” He contends that most hate speech laws forbid only expression that is intended to have the effect of stirring up hatred against a certain group of people. He then notes that because of this intent requirement the “self-same proposition” that would violate a hate speech law would not come within its prohibition if uttered without the requisite intent or under circumstances in which the speaker could not reasonably foresee a prohibited effect. For this reason, Waldron concludes, hate speech laws “get at content only by virtue of its intended effect on the community, rather than on the sole basis of the propositions expressed” and therefore are only “sort of content-based.” 
Waldron’s complicated argument is easily refuted. It is the speech that a law prohibits, not that which it leaves unregulated, that is relevant to determining whether a law is content-based or content-neutral. To use one of Waldron’s own examples, a law that prohibits someone from shouting “Fuck!” in public is no less content based because it does not prohibit people from yelling this “self-same” profanity in private. Waldron, is, of course free to modify his previous conclusion that hate speech bans are undoubtedly content based. He cannot, however, avoid the persuasiveness of his previous analysis or escape his apt observation that arguments such as the one he now deploys are “casuistry.”
Waldron is on somewhat firmer ground in arguing that the requirement that hateful ideas “be expressed in a certain manner,” such as by the use of “threatening, abusive or insulting” language as required by section 18(1) of the UK’s Public Order Act, does not prohibit “certain views per se.” In theory at least a ban on highly vituperative expression might still allow a speaker to convey “something like . . . the propositional content” of the view he wants to express. So upon further reflection, I agree with Waldron that in the functional sense I am primarily using the term viewpoint discrimination in this discussion, that is, preventing a speaker from expressing a particular view in public discourse, hate speech laws that do not significantly impair a speaker’s ability to express bigoted views in public discourse are not viewpoint discriminatory. For if a speaker can still express the basic propositional content of the idea he wants to convey, the regulation would have at most only minimal implications for the basic democratic precept of formal equality and hence for normative legitimacy. But as I demonstrated in my Opening Article and discuss below, Waldron’s assertion that “most” hate speech laws ban only highly vituperative expression of bigoted views can be seriously doubted. Rather, in many cases hate speech laws have been used to suppress fairly temperate expression. 
Finally, Waldron challenges the emphasis I place on whether speech restriction is viewpoint discriminatory or instead more broadly bans expression in a viewpoint-neutral manner. He points out that under both regulations “[s]omeone is still being prevented from saying what he wants to say as he says it. To that effect there is still an impact on the quality of public debate: it is not as it would be if there were no restrictions.” He therefore asks, “[W]hy should the nature of the restriction—viewpoint-based or non-viewpoint-based—make all the difference here?”
The answer to this query is the same as I gave in distinguishing his scenario in which someone was disenfranchised by onerous ID-requirements or stringent restrictions on early voting: viewpoint-discriminatory speech restrictions implicate the core precept of equal citizenship in a way that non-discriminatory curtailment of political participation does not. Compare, for instance, a law that forbids anyone from protesting on the sidewalk in front of the United States Supreme Court with one allowing protests except for those against the Court’s decisions legalizing abortion. Under both laws anti-abortion demonstrators are “being prevented from saying what [they] want to say . . .” But it cannot be seriously gainsaid that the viewpoint-based law infringes core democratic equality concerns far more than the more extensive yet viewpoint-neutral law.
- The Extent of Speech Curtailment
In my Opening Article I discussed a number of cases in which people were arrested, prosecuted or convicted under hate speech laws for relatively temperate expression of racist or anti-Islamic speech or for criticizing homosexuality as immoral or disordered. Among the cases I discussed were the conviction of a Dutch politician for calling for the removal of “all Surinamers, Turks and other so-called guest-workers from the Netherlands”; the fining of a an Austrian academic for saying that Mohammad “had a thing for little girls”; the fining of actress Brigitte Bardot for writing that Muslims were destroying France by “imposing their ways”; and the unsuccessful prosecution of the Catholic Bishop of Belgium for expressing the view that homosexuality was a “blockage in normal psychological development.” I claimed that these and other cases show that contrary to Waldron’s assertion, hate speech bans “manifestly do not in practice provide a ‘safe haven’ for expressing ‘something like the propositional content’ of bigoted views that become illegal only ‘when expressed as vituperation.’” I acknowledged that such applications to temperate speech do not “represent the typical hate speech case,” which often does involve vituperation. I emphasized, however, that in evaluating these applications to non-vituperative speech, we need to consider not just the effect on the speakers in these cases but also the “chilling effect” these cases would likely have on would-be speakers.
In her short yet incisive Commentary, Gelber insists that the thirty-two cases that I discuss are insufficient in “empirical terms” to disprove Waldron’s contention that most hate speech laws provide a “safe haven” for the temperate expression of the basic “propositional content” of the view the speaker wants to express. She insists that a “far more in-depth and systemic study would be needed regarding the operation of hate speech law in practice to sustain this point.” Gelber’s criticism is well taken. What I should have said is that these cases cast considerable doubt on the validity of Waldron’s assertion that “most [hate] speech laws bend over backward to ensure that there is a lawful way of expressing something like the propositional content of views that become objectionable when expressed as vituperation.” Or even if Waldron is right that the intent of most such laws is to create such “safe havens” for non-vituperative expression of bigoted views, the cases I discuss raise serious concerns whether in practice most hate speech laws have actually created such safe zones.
I am not sure if Gelber would concur with even these modified claims. It is interesting to note though that she does agree that some of the cases I mention “ought not to have been considered hate speech and ought not to have been prosecuted.” Waldron, in contrast, is remarkably unconcerned by the cases I discuss in which no such “safe harbor” was provided for non-vituperative expression of bigoted views. The entirety of his response to my examples of fairly temperate speech having been prosecuted is this: “No doubt there are hate speech laws in the world expressed less carefully, with less attention to these fastidious distinctions than the British provisions I have cited. But our debate is about hate speech restrictions as such, not about the least well-formulated of them.” Professor Blasi is surely correct, therefore, when he observes that “Waldron fails to grapple as fully as he needs to with the challenges to his argument” that these cases pose.
Gelber also is correct in claiming that there is “insufficient evidence” to support my claim that the cases I discuss “undoubtedly” have had a “chilling effect” on non-vituperative speech. Although United States Supreme Court decisions frequently rely on such an effect, there is surprisingly very little empirical support for the existence of this phenomenon. Significantly, however, even if the uncertain reach of many hate speech laws does not deter people from expressing views that the laws did not in fact mean to ban, it would still be the case that in many jurisdictions someone wanting to call for the removal of “so-called guest workers” of certain ethnicities from their country; or to condemn Mohammad as a pedophile; or argue that Muslims were destroying the country by “imposing their ways”; or to claim that homosexuality was a “blockage in normal psychological development”; would have reasonable grounds to fear arrest for doing so. For here it is not just the uncertain reach of the laws that may (or may not) “chill” such speech. Rather, the deterrence arises from the fact that others have been arrested, and in many cases convicted, for conveying precisely the views the speaker wants to express.
There is also merit to Gelber’s objection, echoing Waldron’s complaint, about my use of cases decided under Section 5 of the UK’s Public Order Act to show that there was no “safe haven” for relatively temperate expression of expression condemning homosexuality or Islam. Gelber writes that if I wanted to show “the misapplication of hate speech laws” that it would have been helpful if I had limited my inquiry to these laws. I invoked the Section 5 cases, however, not to bolster my claim about misapplication of hate speech laws per se but rather to show that even where, as in the UK, the hate speech laws might intend to create a safe haven for relatively temperate expression, other laws might still prohibit such expression. But this point would have been clearer if I had considered the Section 5 cases separately and not interspersed them with misapplication of hate speech laws.
Still, Waldron is wrong when he says that these Section 5 cases have “nothing to do with our disagreement about hate speech.” To the contrary, as Gelber correctly notes, Section 5 has been “used to shut down what was perceived to be hate speech.” For instance, in determining that a street preacher’s placard referring to homosexuality as immoral violated Section 5, the Court of Appeal emphasized that “there is a need to show tolerance towards all sections of society” and that therefore the message on the preacher’s placard “went beyond legitimate protest.” The Norwood case even more directly involves expression punished because it was “perceived to be hate speech.” In that case, the penalty for non-vituperative speech insulting to Islam (which the Court of Appeal also condemned as going “beyond legitimate protest”) was enhanced because the expression was “motivated (wholly or partly) by hostility towards members of a racial or religious group based on their membership in that group.”
It is true that the Section 5 cases I discussed do not directly impugn the operation of hate speech laws in Britain or elsewhere. These cases do, however, put into doubt any claim that there is a “safe haven” even in Britain for non-vituperative speech critical of homosexuality or insulting to Islam. Moreover, even though the decisions upholding convictions for this expression may not have expressly used the term “hate speech,” this is precisely the type of expression prohibited by hate speech laws in other democracies. In the United States, provisions against disorderly conduct are routinely used to punish speech, including “fighting words” and true threats. But precisely because in the United States hate speech is not a category of expression that can be constitutionally suppressed, public order acts in the United States cannot be used to punish hate speech as such but only if it otherwise constitutes unprotected speech.
To conclude this discussion of the extent of the actual curtailment on expression imposed by hate speech requirements, I want to return to the Evangelical Photographer Scenario. In my Opening Article I contended that in many jurisdictions with hate speech laws on the books someone might reasonably fear being arrested, and perhaps even convicted, for expressing the view in the public square that homosexuality was immoral. So while my “anecdotal claims” about how hate speech laws actually operate might not have with adequate social scientific rigor disproved Waldron’s far less empirically-supported assertion about “safe havens” for temperate expression; and while there may be little empirical evidence that hate speech laws “chill” speech not intended to be proscribed; it is telling that no participant in this Symposium denied my claim that it would be reasonable for a citizen in many democratic jurisdictions with hate speech bans to forego even such relatively temperate criticism of homosexuality for fear of arrest. Even in the absence of definite proof, the distinct possibility that relatively temperate speech in the public square opposing the addition of sexual orientation to a nation’s antidiscrimination laws, or criticizing Islam as part of opposition to a country’s immigration policy, should raise real concern about the potential of hate speech laws to undermine the legitimacy of downstream antidiscrimination laws.
In response, several commentators argue that legitimacy concerns appear on both sides of the equation in that hate speech itself can impair political legitimacy by curtailing the ability of the subjects of such speech to participate as equals in the political process. I agree that any theory grounding free speech in political legitimacy must carefully consider any countervailing legitimacy interests for speech restrictions. For this reason, I raised and discussed this issue in my Opening Article and will now respond at length to the various arguments made by Commentators that hate speech itself can impair legitimacy.
- Countervailing Legitimacy Concerns and the “Silencing” Argument
Proponents of hate speech laws often claim that hate speech silences its victims. For anyone concerned about political legitimacy, this claim must be taken seriously. People being prevented from participating in public discourse for any reason, including by the speech of others, will interfere with the ability of those who have been silenced “identify[ing] with the state in the manner required by democratic legitimacy.” In addition, as I explained in my Opening Article, such silencing might render immoral the enforcement of particular laws against those whose speech has been curtailed. The silencing argument, however, suffers from a crucial defect: although the claim has been made for decades, there is to date at most only meagre evidence that hate speech as part of public discourse—and it is only bans on hate speech as part of public discourse that I argue impair political legitimacy—prevents anyone from “contribut[ing] to public discourse [and] participat[ing] in the formation of public opinion.” It is important, therefore, to see various silencing arguments proffered in this Symposium for what they really are: studious attempts to avoid the problem that there is a “paucity of evidence” supporting this argument.
- Silencing, Legitimacy and Hypothetical Consent
Alexander Brown argues that the “assurance of civic dignity” that Waldron emphasizes is “constitutive of the realization of political legitimacy.” In Brown’s view “political legitimacy, including the legitimacy of the legal system, itself depends upon it being possible, at least in principle, to justify that system to each citizen bound by it on the basis of fundamentals of justice that they cannot reasonably reject.” Following this line of reasoning Brown asks:
[W]ould free and equal people have reasons based on the fundamentals of justice to reject an aggressive free speech regime that treated hate speech as a protected category even though certain forms of hate speech carry a risk of effectively removing from some people who are the subject of hate speech real opportunities to contribute to public discourse and participate in the formation of public opinion?
Brown does not specifically answer this question. But his answers to previous iterations of his “process of interpersonal justification and consensus” leave little doubt that he thinks the answer would be “yes.”
I agree that it might be reasonable for “members of minority or vulnerable groups” to reasonably reject as contrary to “fundamentals of justice” an “aggressive free speech regime” such as exists in the United States. But the reasonableness of this conclusion would depend on such issues as: 1) the types and extent of hate speech currently prevalent in the society in question; 2) the extent to which these various forms of hate speech “carry a risk” of preventing the subjects of hate speech from participating in public discourse; 3) whether the reaction to the hate speech was reasonable; 4) causes other than hate speech for this “silencing” effect; 5) the likelihood that hate speech bans will remedy this harm; 6) the availability and likely effectiveness of viewpoint-neutral speech restriction, such as general bans on threats, fighting words and incitement to violence; 7) the availability and likely effectiveness of non-coercive alternative remedies; 8) the extent to which the proposed hate speech laws are likely to be misapplied and thus might silence people, including members of “minorities or vulnerable groups” from participating in public discourse; and 9) the benefits that all members of society, including “minorities and vulnerable groups derive” from a “free speech regime” that “aggressively” protects against viewpoint discrimination.
The answers to these questions will, of course, vary from society to society as well as over time within each society. But without careful consideration of these and other largely empirical questions, it is impossible to make any meaningful judgment about whether it is reasonable to reject the failure to enact certain types of hate speech law as contrary to the “fundamental justice” required for people being willing to join a political community. For this reason, Brown’s “process of interpersonal justification and consensus” is incapable of doing anything more than showing that “under certain circumstances” hate speech can have a detrimental impact on political legitimacy, either through “silencing” or by inflicting other harms, a proposition with which I readily agree.
Significantly, and consistent with this conclusion, Brown’s “process of interpersonal justification and consensus” could be employed to support an “aggressive” free speech regime that would prevent the silencing of legitimate (and legitimating) dissent. According to Brown, fundamentals of justice include “everyone’s claim to have their welfare counted along with everyone else’s welfare in the determination of social policy.” It could therefore be argued that those who hold traditional religious beliefs (among others who may want to vigorously challenge current political orthodoxy) might reasonably reject as contrary to fundamental justice any legal regime that permitted viewpoint-based restrictions on public discourse, including hate speech bans, on the grounds that even if narrowly drafted such laws would in actual practice present “a risk of effectively removing from some people . . . real opportunities to contribute to public discourse and participate in the formation of public opinion.” As with Brown’s use of the process of interpersonal justification and consensus to argue against “aggressive” free speech regimes, whether rejection of hate speech bans as contrary to fundamentals of justice is reasonable would depend on a host of empirical inquiries including the scope of the speech restriction and whether it would prevent even temperate criticism of homosexuality (or some other currently dominant opinion in society). So, all that “the process of interpersonal justification and consensus” is capable of doing is showing that both hate speech bans and hate speech itself can potentially damage political legitimacy, a proposition that I expressly acknowledged in my Opening Article.
Finally, even if in a particular mature, stable democratic society it was reasonable for “minority and vulnerable groups” to reject the failure to enact a particular type of hate speech laws as contrary to “the fundamentals of justice,” difficult empirical (as well as normative) questions would remain as to the extent of the damage the failure to enact these types of hate speech laws has inflicted on the legal system. And precisely because legitimacy is potentially on both sides of the equation, this reduction in legitimacy would then have to be balanced against the damage to the legitimacy of the legal system that hate speech bans would likely cause.
- Silencing and the Precautionary Principle
In my Opening Article I remarked that Brown’s invocation of the precautionary principle to compensate for the lack of evidence supporting the silencing argument turned “a problematic though plausible argument into a plainly indefensible one.” In his Commentary, Brown continues to try to defend the indefensible. He begins by discussing arguments for banning hate speech as a prophylactic measure against genocide or “terrorist atrocities” that threaten national security. He then refers to these “grave and irreversible” harms as “equivalent to the devastating climate change harms that are associated with the precautionary principle in the field of environmental regulation.” But having invoked these catastrophic harms, Brown then concedes that none of the harms that he argues justify suppression of hate speech, including its alleged silencing effect, is “of the same magnitude of gravity” as genocide, terroristic attacks threatening national security or climate change. He insists, though, that harms from hate speech that he thinks justify its suppression, including the silencing effect, “are potentially more probable harms and more proximate harms, causally speaking.” It is the greater probability and closer causal connection between hate speech and harm that in Brown’s view put the silencing effect on par with genocide, terroristic atrocities and global warming so far as invocation of the precautionary principle is concerned. Accordingly, he concludes that these factors justify putting a heavy burden of proof on those who oppose hate speech bans to show that the silencing effect will not occur if hate speech is left unregulated.
There are at least four salient defects with this remarkable argument. First, Brown does not claim that on the scale of gravity of harm, the silencing effect even approaches genocide, terroristic atrocities or global warming. So invocation of these truly catastrophic harms is largely a distraction. In addition, Brown’s argument is question begging. That hate speech is potentially more likely and proximately to deter people from participating in public discourse than it is to lead to genocide or to threaten national security tells us next to nothing about how likely in any given society hate speech is to deter people from participating in public discourse. But it is this probability that Brown relies on to justify applying the precautionary principle.
The most remarkable problem with Brown’s argument, however, is the dearth of evidence that he contends justifies applying the precautionary principle to impose a particularly heavy burden of proof on those who oppose hate speech bans. To trigger application of the precautionary principle, proponents of hate speech bans would have to adduce only “some minimally adequate evidence” of the silencing effect. It is one thing to invoke the precautionary principle to avoid a catastrophic or perhaps even a non-catastrophic though serious harm in the absence of “scientific certainty” that the harm will occur. It is quite another to invoke the principle when the evidence of a serious but non-catastrophic harm can meet the vague “some minimally adequate evidence” standard. Use of this heretofore unknown standard to trigger the precautionary principle seems especially concocted to fit the meagre evidence at hand.
Finally, and relatedly, Brown fails to take account of the risks of applying the precautionary principle to ban hate speech arising from the misapplications of these laws that I have discussed. As Professor Cass Sunstein has explained, any application of the precautionary principle that does not consider “the risks on both sides of a decision” is “deeply incoherent.” Indeed, if as Brown argues probability and proximate cause are to be given the decisive weight in supporting the propriety of applying the precautionary principle, a better case might be made that the principle supports not enacting hate speech bans. In light of the meagre evidence that has been produced in support of the silencing effect in most mature and stable democracies, it could be strongly argued that imposing criminal sanctions for expressing hateful views in public discourse would more likely and more proximately silence people from expressing their views in public discourse than does hate speech itself.
In the course of making this remarkably weak argument for application of the precautionary principle Brown does, however, give a plausible answer to my longstanding query of what principle would “justify shutting up A (or a group of As) so that B (or a group of Bs) can speak?” Brown argues that criminalizing the types of hate speech that he identifies for suppression will “will not stop the speaker from expressing him or herself in other permissible ways.” In contrast, he continues, “in the event that hate speech has a silencing effect on those who are its subjects, the effect is just that, silence; it can cause people not to speak in any way.” Of course, hate speech will not (unless it causes death or a catatonic state) literally “cause people not to speak in any way.” It might conceivably, though, even in a mature and stable democracy prevent people from engaging in public discourse not just on particular subjects, but entirely. Such an asymmetry between the extent of silencing caused by hate speech and hate speech bans might depending on the size of each group provide a principle for shutting up a group of As so that a group of Bs can speak. Indeed, if there was persuasive evidence that hate speech in a certain society was preventing a large group of people from participating in public discourse entirely, it might be appropriate to invoke the precautionary principle in support of hate speech laws even in the absence of “scientific certainty” that such harm was occurring or was imminent. But evidence for such a “catastrophic antidemocratic outcome” is even more wanting than it is for the usual silencing argument.
In concluding this discussion of Brown’s use of the precautionary principle, I want to take a page out of Waldron’s book (or rather some phrases from his Response): I wonder if such a strong, idiosyncratic, empirically-insensitive version of the precautionary principle is not “just being wheeled into the hate speech debate” as a theory “rigged up for the purpose of the hate speech debate.” As Waldron correctly notes, one way of showing that a principle is “not rigged” is to follow it where it leads. So I am curious whether Brown would be willing to apply this version of the precautionary principle to suppress anti-war protests if a couple of studies provided not particularly relevant evidence that such protests might lead to significant harms to a war effort, such as interference with recruitment of personnel. In this regard, it should be noted that these harms are “potentially more probable harms and more proximate harms, causally speaking” than graver harms that possibly could result from anti-war protests, such as increased battlefield casualities and even the loss of the war resulting from expression that dispirits our troops and encourages the enemy.
- Silencing and the Burden of Proof
Rather than trying to shift the burden of proof to those who oppose hate speech laws as does Brown, Adrienne Stone makes the more modest claim that “there is no reason, in the absence of evidence on silencing, to err on the side of free speech.” There is, in fact, a very powerful, and it seems to me conclusive, reason for erring on the side of free speech. Or more precisely, there is a good reason in our present state of knowledge for not enacting hate speech laws if making sure that people are not prevented from asserting certain views in public discourse is the goal. It is indisputable that hate speech bans prevent people from expressing certain views in public discourse. Silencing the expression of certain views is after all is the purpose of such laws. In contrast, as Gelber concedes, there is “not a great deal” of evidence that hate speech causes such silencing. In any event, Stone’s claim about on which side we should err assumes for the sake of argument that there is not just meagre evidence of the silencing effect but rather an “absence of evidence on silencing.” If and when there is persuasive evidence that hate speech as part of public discourse actually prevents members of minority or other vulnerable groups from expressing some viewpoint in public discourse, or worse yet, prevents them from participating in such discourse altogether, a very difficult problem will be presented. For under these circumstances it will have to be determined whether a greater degree of silencing, and hence potential damage to political legitimacy, is likely to be produced by hate speech bans or by hate speech itself. But in the absence of any evidence of the silencing effect of hate speech, or even with what is at best meagre evidence of such an effect that has been proffered to date, we should indeed “err on the side of freedom of speech” so far as avoiding damage to political legitimacy is concerned.
- Silencing and Speech-Act Theory
Relying on an influential essay by Professor Rae Langton, Gelber and Stone argue that hate speech might silence members of historically-oppressed groups, not only literally by keeping them from speaking as Brown alleges, but also by disabling them from conveying the meaning their words ordinarily would import. Langton argues that the way women are portrayed in pornography might have such a disabling effect on a woman’s ability to say “no” to man’s sexual advances. On this view, a man who consumes pornography might simply not understand what the woman is trying to convey by saying “no,” thereby rendering her ability to convey her lack of consent “unspeakable.” It may be that pornography has some role in contributing to the deranged belief of too many men even in modern democracies that when a woman says “no” to a sexual advance she does not mean “no.” Neither Gelber nor Stone, however, suggests an equivalent form of silencing arising from hate speech. Significantly, they do not even speculate on how hateful expression as part of public discourse might deprive a participant in public discourse of the ability to convey her message in a way analogous to pornography’s purported silencing effect on a woman trying to verbally refuse a sexual advance.
Rather, Gelber and Stone, in addition to discussing the literal silencing effect claimed by Brown, may be arguing that vilification of members of historically-oppressed groups can impair the ability of those so villified to effectively convey their views. I agree that it is likely that those persuaded by hate speech defaming, for instance, African-Americans, Muslims, or Jews, would tend to discount anything members of these groups have to say on matters of public concern. But while it is important to recognize this unhappy consequence of hate speech, preventing this harm is, in my view, not an appropriate reason for suppressing expression in any society in which “public opinion . . . is the final source of government.” Unfair characterization of people, or groups of people, likely to have such an effect is by no means confined to hate speech. Rather, it is a regrettable feature of public debate on virtually any subject but especially with discussion of contentious, highly-ideological issues. It would be difficult to find a principled ground for limiting this variation of the silencing argument to hate speech. Without such a limitation, however, it would provide a rationale for government to suppress a large swath of public discourse that must be allowed in a free and democratic society.
- Hate Speech as Constituting Harm
Again relying on Langton, Gelber in several places in her Commentary insists that hate speech does not just lead to harm but by its very utterance constitutes harm. For instance, Gelber writes that “the defining features of hate speech are . . . that it incurs harms discursively when the hate speech is uttered, and that these harms are analogous to other discriminatory harms, such as denying someone a service or denying them a job on the ground of their race.” Significantly, however, Gelber does not explain or give an example of how hate speech in public discourse can constitute a harm analogous to the ones she mentions. I can easily imagine how speech outside of public discourse, for instance an employer calling his black employees “niggers,” or a “Whites Only” sign on a the door of a restaurant, could be said to constitute harm rather than just lead to race discrimination in employment or public accommodation. In contrast, while I agree that hate speech within public discourse might lead to discrimination against members of vulnerable minority groups, I am unable to think of a plausible example of how such speech constitutes harm analogous to the examples just mentioned.
- THE SEARCH FOR A CABINABLE RATIONALE SUPPORTING HATE SPEECH BANS
A daunting challenge for those supporting hate speech bans is to find a rationale for excluding hate speech that would also not justify suppression of speech they believe should not be banned in a free and democratic society. I am not denying that such a principle exists. Indeed, the justification of the suppression of hate speech on the grounds that it prevents others from expressing views in public discourse or, worse yet, from participating in this public discussion altogether, might provide such a rationale. The problem with this silencing argument is not conceptual but empirical, for as we have seen, there is scant evidence that hate speech has such a silencing effect in any mature and stable democracy. I have already alluded to the untenable scope of various rationales that participants in this Symposium have offered to try to obviate this glaring evidentiary problem. But the problem of overly broad rationales for suppressing hate speech is sufficiently troubling to warrant a more focused discussion.
Proponents of hate speech bans, including several participants in this Symposium, make arguments that would also apply to speech that I am confident they believe must be allowed in a free and democratic society. I have already discussed how under Brown’s version of the precautionary principle anti-war speech could readily be banned. And so too under this rationale, government could ban the mere glorification of terrorism or advocacy of anti-democratic forms of government, including communism. Similarly, Stone’s view that there is no reason “to err on the side of free speech, in the absence of evidence of silencing” would apply to suppressing speech to prevent other serious harms even without evidence of their existence.
In the much the same way, the speech-act theory that Stone and Gelber invoke in support of suppressing hate speech could readily be applied to speech that I believe they would not agree could be properly suppressed. I have already discussed how one particular version of a silencing argument would justify suppressing any view in public discourse that arguably has that effect. In addition, like Brown’s precautionary principle, their argument that hate speech in public discourse can be banned because it not just causes but “constitutes” discrimination could be invoked to suppress anti-war speech condemning the draft as “constituting” draft resistance. Finally, even Shiffrin’s nuanced, harm-based rationale for suppressing hate suffers from this overbreadth problem.
According to Shiffrin, hate speech causes the following harms:
it is an assault on the dignity of people of color; it humiliates and causes emotional distress, sometimes with physical manifestations; it helps spread racial prejudice, not only stigmatizing people of color in the eyes of the societally dominant race but also in the eyes of [many of] the victims themselves, inspiring self-hatred, isolation, and . . . finally, it frequently creates the conditions for violence.
Although Shiffrin does not maintain that these harms are sufficient to suppress full-value public discourse, he does insist that they are sufficient to suppress hate speech because such expression ranks “low in the hierarchy of First Amendment values.” What makes hate speech “low value” expression in Shiffrin’s view is that in addition to being harmful, those engaging in such expression “seek to topple the fundamental prerequisites of a legitimate society and government,” including “the system’s foundational premise of equality.” The problem with this “low value” speech rationale is that it can be used to suppress advocacy of radical political change, including certain forms of Marxist speech. Like racist speech, advocacy of the violent overthrow of democratic institutions can “create the conditions for violence.” And since advocacy of violent overthrow of democratic governemnt seeks “to topple the fundamental prerequisites of a legitimate society and government,” such expression can on Shiffrin’s account be deemed “low value” expression suppressible even in the absence of the showing of harm needed to suppress full-value speech. Indeed, it is precisely this rationale that a plurality of the United States Supreme Court used to uphold the conviction of high ranking members of the American Communist Party for violation of the Smith Act.
It is worth emphasizing that I am not here making a “slippery slope” argument. Unlike a slippery slope rationale, my claim says nothing about what other forms of speech government in fact will or is even likely to suppress if allowed to suppress hate speech. Rather, I am claiming that expression that proponents of hate speech bans likely would want protected is fairly encompassed within various rationales they proffer to support the suppression of hate speech. If I am right about this, then those who have offered such an overly capacious rationale should consider retracting it, not because it might in fact lead via a slippery slope to the suppression of speech that they think should be protected, but because it shows that the rationale is faulty. Relatedly, these rationales should also be rejected because they unduly weaken a democratic nation’s free speech principle.
- Expanding the Inquiry
Robert Post’s Commentary provides a useful supplement to my analysis. Although briefly discussing how viewpoint-based speech restrictions, including hate speech bans, can damage systemic legitimacy, the burden of my Opening Article was to explore the potential of hate speech bans to damage the normative legitimacy of particular downstream laws. Robert Post’s Commentary, in contrast, focuses on how speech restrictions might impair the descriptive legitimacy of the legal system. In this way, Post usefully expands the discussion.
I agree with Post that freedom of speech “allow[s] persons of widely varying views to experience as legitimate a government that may nevertheless act in ways that are inconsistent with their own ideas” and that if “persons are prevented from expressing their own views—however much others might find those views outrageous and intolerable—then they are less likely to experience their government as legitimate.” I also agree with him that hate speech bans restricting the abusive manner in which bigoted ideas can be expressed in public discourse can significantly damage systemic legitimacy in the descriptive sense by alienating people from the legal system. But this is not primarily, as Post suggests, because such restrictions even if properly applied to punish only bigoted ideas expressed in a higly abusive manner, would interfere with these speakers’ ability to influence “the shape of public opinion” to which “their representatives are supposed to be responsive.” As I discuss above, I think that Post overestimates the extent to which limitations on “abusive language” interfere with speakers’ ability to express the precise idea they wish to disseminate. Rather, it is primarily the selective nature of these prohibitions that in my view is likely to damage descriptive legitimacy. This is because limitations on the abusive manner of expression applicable to hate speech, but not to other types of abusive expression in public discourse, are likely to be perceived by bigots as an unfair attempt to muzzle them, but not opponents of equality. In addition, as the United States Supreme Court has recognized, speech has an emotive as well as a cognitive function. Accordingly, regardless of any impairment that they may have on speakers’ ability to “influence the shape of public opinion,” laws that prohibit participants in public discourse from employing abusive language to vehemently express a viewpoint might interfere with speakers’ “experienc[ing] as legitimate a government that might nevertheless act in ways that are inconsistent with their own ideas.”
Post and I are in accord that “democratic legitimacy is at stake on both sides of the equation” in that such legitimacy is adversely affected by the simple act of prohibiting hate speech but depending on “the particularities of national circumstances” legitimacy can also be undermined by allowing hate speech . . . .” But it may be that Post’s view of the potential of hate speech to damage descriptive political legitimacy is somewhat greater than I conceive it to be. In Post’s view descriptive legitimacy concerns the “conditions necessary for a diverse and heterogeneous population to live together in a relatively peaceable manner under a common system of governance and politics.” On this specification, expression that impedes the goal of such a population “liv[ing] together in a relatively peaceable manner,” an effect that hate speech might well have in various societies, would damage legitimacy even if it did not prevent anyone from participating in the democratic process. Such intergroup conflict is a serious harm that should be accounted for in any assessment of the propriety of hate speech laws in a free and democratic society. I am not sure, however, it should be counted as a detriment to political legitimacy, at least not without further explanation.
Vincent Blasi’s superb Commentary begins by meticulously, accurately and fairly summarizing the essence of both Waldron’s and my positions in our debate about hate speech regulation and political legitimacy. It then points out what he considers to be the strengths and weaknesses of both of our positions. I am enormously grateful to Blasi for taking the time and care to understand what I was trying to accomplish in my Opening Article.
Blasi kindly credits my argument as offering a “fresh account of why viewpoint discrimination might be problematic” in arguing that viewpoint-discriminatory restrictions on public discourse offend the basic democratic premise that each person in society has “equal civic standing.” He notes that on this account “viewpoint discrimination is not about the quality of public debate,” as posited by Professor Geoffrey Stone, or about “the rightful sources of governmental authority,” as emphasized by Judge Learned Hand. Rather, Blasi correctly observes that my indictment of viewpoint discrimination derives from “the equal treatment of individuals,” which entails “[r]espect for the civic dignity of each individual speaker.”
Blasi accurately notes that any answer I might have to Waldron’s “powerful critique” of my claim that upstream hate speech restrictions can deprive downstream antidiscrimination laws of legitimacy has to come from my equality-based rationale for impugning viewpoint discrimination. He comes to no conclusion, however, about whether this rationale provides sufficient support for my “imaginative legitimation argument, which Waldron has questioned effectively.” Rather, he thinks I have a “more straightforward line of support” from my case against hate speech regulation, namely, an argument based directly on equal civic standing. Such an argument, in Blasi’s view, would have the advantage of being commensurable with Waldron’s argument that hate speech undercuts the civic dignity of members of vulnerable minority groups.
Blasi does not decide whether the insult to civic dignity wrought by hate speech bans is justified by the protection of civic dignity it produces but rather offers some observations about the “variables” that might be used in such an analysis. Except for some reservations about his claim that “the most important value at stake in the comparison” is freedom of thought, I agree that the approach Blasi suggests would be fruitful. But more importantly, I applaud Blasi for building on what he generously referred to as a “promising line of inquiry” opened up in the debate between Waldron and me to expand the discussion in a useful direction.
It has long been recognized that free speech promotes the legitimacy of a legal system. This is one important reason that free speech is highly valued in liberal democracies. It also explains why a proliferation of laws that impede the ability of citizens to express their views on matters of public concern would substantially impair the legitimacy of a legal system. But as crucial as free speech is to political legitimacy, it cannot plausibly be argued that even the most egregious law restricting free speech that realistically might be enacted in a mature and stable democracy could destroy or even severely damage the legitimacy of the entire legal system. Rather, given the level of legitimacy that these democracies enjoy, even the most poorly-justified speech restriction likely to be enacted will usually only reduce the legal system’s legitimacy “reservoir” by some barely perceptible amount. For this reason, those who favor restraining speech that they believe inimical to some goal, policy or value that they care passionately about will likely conclude, at least so far as legitimacy is concerned, that banning such speech is well worth the miniscule reduction in the legal system’s ample legitimacy reserve.
In contrast, unlike the negligible damage that a single speech restriction can inflict on the systemic legitimacy of a mature and stable democracy, the harm to the legitimacy of particular downstream laws resulting from an upstream speech restriction can be considerable, sometimes even ruinous. Of course, such damage to legitimacy, both descriptive and normative, will be more likely if the speech restriction is dramatically lacking in justification. But as my Evangelical Photographer scenario and some actual cases I discuss reveal, this palpable damage to legitimacy can sometimes result even if the speech restriction is, as I believe is the case with many hate speech laws on the books, not “indefensible” but merely on balance unjustified.
So did this novel take on the relationship between free speech and political legitimacy, developed beyond cursory statements for the first time in this Symposium, survive the intense scrutiny to which it was subjected by the Response and eight Commentaries? To my mind, the most powerful objection was the contention that if a speech restriction was justified, it could not deprive a downstream law of normative legitimacy. As I explained, I am not sure that this objection is valid, especially when the iterative nature of a legal justification is considered. But even if the objection is correct, the idea I defended nonetheless serves to reveal a previously unidentified cost of unjustified speech restrictions. In addition, this objection has no bearing on the damage to descriptive legitimacy that can result from even justified viewpoint-discriminatory speech restrictions.
In contrast, the other objections did not take issue with my basic contention that speech restrictions that unduly prevent people from expressing their views in public discourse can compromise, both normatively and descriptively, the legitimacy of downstream legislation. Rather, this criticism focused on my claim about the scope and extent of the impediment imposed by hate speech bans on the expression of bigoted viewpoints in public discourse; on my assessment of the potential for hate speech itself to undermine political legitimacy; and on the proper legal response in the absence of any persuasive evidence that hate speech prevents the subjects of such speech from participating in the political process. I am content to leave open for now the question of whether hate speech laws as they currently operate in many democracies damage the legitimacy of downstream laws, particularly antidiscrimination measures, and if so, the extent to which this loss of legitimacy might be offset by increased participation in public discourse by the subjects of hate speech. What I hope that I have demonstrated in this Symposium is that a law that restricts the right of citizens to express particular views in public discourse—be it a hate speech ban, a blasphemy law, a prohibition on anti-war speech, or a law that forbids advocacy of radical political change—can impair, and in rare cases even destroy, the legitimacy of downstream laws.
* Dan Cracchiolo Chair in Constitutional Law, Sandra Day O’Connor College of Law, Arizona State University. I am grateful to Trevor Allan, Dick Arenson, Jill Hasday, Jeff Murphy, Robert Post, Jake Rowbottom, Fred Schauer, Mary Sigler, Cynthia Stonnington, Wayne Sumner for their helpful comments and to James Hall, Luci Davis, and Austin Yost for their valuable research assistance.
. Vincent Blasi, Hate Speech, Public Assurance, and the Civic Standing of Speakers and Victims, 32 Const. Comment. 585 (2017); Alexander Brown, Hate Speech Laws, Legitimacy, and Precaution: A Reply to James Weinstein, 32 Const. Comment. 599 (2017); Katharine Gelber, Hate Speech—Definitions and Empirical Evidence, 32 Const. Comment. 619 (2017); Eric Heinze, Taking Legitimacy Seriously: A Return to Deontology, 32 Const. Comment. 631 (2017); Robert Post, Legitimacy and Hate Speech, 32 Const. Comment. 651 (2017); Frederick Schauer, Free Speech and Obedience to Law, 32 Const. Comment. 661 (2017); Steven H. Shiffrin, Hate Speech, Legitimacy, and the Foundational Principles of Government, 32 Const. Comment. 675 (2017); Adrienne Stone, Viewpoint Discrimination, Hate Speech Laws, and the Double-Sided Nature of Freedom of Speech, 32 Const. Comment. 687 (2017).
. Jeremy Waldron, Hate Speech and Political Legitimacy [hereinafter Waldron, Political Legitimacy], in The Content and Context of Hate Speech 329, 339–40 n.43 (Michael Herz & Peter Molnar eds., 2012); Jeremy Waldron, The Harm in Hate Speech (2012) [hereinafter Waldron, Hate Speech].
. See Blasi, supra note 3, at 585 (In “open[ing] up a promising line of inquiry regarding the legitimacy and propriety of hate speech regulation” this discussion has “succeeded in reinvigorating a subject that [has] grown academically formulaic even while becoming alarmingly more salient politically and culturally.”); Stone, supra note 3, at 687 (“The essays to which we are responding take the long and rather well-worn debate about hate speech in new directions.”).
. Waldron is mistaken that I agree with what he characterizes as Dworkin’s concession that the potential effect of a speech restriction on legitimacy “was diminution rather than destruction” of downstream legislation. Waldron, supra note 2, at 705. First of all, Waldron may be over reading the extent of Dworkin’s concession. In agreeing that, despite the restrictions on political expression imposed by its upstream restrictions on hate speech, “[o]n balance Britain is entitled to enforce” its downstream antidiscrimination laws, Waldron, Political Legitimacy, supra note 4, at 335 (quoting email from Ronald Dworkin to Jeremy Waldron, Oct. 4, 2009, 21:34 EST (on file with Waldron)), Dworkin did not necessarily concede that under no circumstances could a speech restriction destroy the legitimacy of a downstream law. Rather, I read Dworkin as leaving this possibility open. In any event, this was precisely the question that I explored in my Opening Article and contrary to Waldron, as shown in my Evangelical Photographer Scenario, Weinstein, supra note 1, at 567, I do believe that under certain circumstances such destruction is possible. Waldron is correct, however, in observing that in my view for a speech restriction to destroy the legitimacy of a downstream law there must already be other difficulties with the morality of enforcing the downstream law. Waldron, supra note 2, at 706. I should add that if Dworkin was referring, as Waldron dubiously claims he was (see infra note 9), to the inability of a single hate speech law to destroy the legitimacy of the entire legal system, then I, of course, agree that it cannot. See text accompanying note 43, infra.
. Until reading Waldron’s response, I thought my position regarding the effect that speech restrictions might have on political legitimacy was essentially the same as Dworkin’s. Waldron, however, considers my position in two crucial respects more “modest” and “focused” than Dworkin’s and thus more defensible. Waldron, supra note 2, at 705–06. First, he thinks that Dworkin was concerned with the effect of hate speech restrictions not on particular downstream legislation, as I am, but with its effect on systemic legitimacy. Id. at 705, Professor Eric Heinze shares this view. See Eric Heinze, Hate Speech and Democratic Citizenship 86 (2016). In addition, Waldron thinks that it was my innovation to focus on the effects of speech restrictions on downstream legislation only with regard to those constrained by the upstream speech restriction. Waldron, supra note 2, at 706. As to his first point, I am not at all sure that Waldron and Heinze are right that Dworkin was concerned with the effect of hate speech laws on the entire legal system rather than their effect on particular downstream laws. Thus Dworkin writes: “[But] if we intervene too soon in the process through which collective opinion is formed, we spoil the only democratic justification we have for insisting that everyone obey these laws.” Ronald Dworkin, Foreword to Ronald Dworkin, Extreme Speech and Democracy vii (Ivan Hare & James Weinstein eds., 2009) (emphasis added). In addition, Dworkin states that “on balance Britain is entitled to enforce such laws.” Id. (emphasis added). As to his second claim, although I was unaware of it, my view that the legitimacy defect in enforcing the downstream law is limited to those constrained by the law may have been a modification of Dworkin’s more encompassing view. Waldron suggests that I made these modifications of Dworkin’s position “simply to make it come out as less implausible than Dworkin’s wholesale version.” Waldron, supra note 2, at 707 Though I did not consciously modify Dworkin’s position for this or any other reason, I would have thought that modifying a position to try to make it “less implausible” is a good thing.
. Waldron, supra note 2, at 709. Especially phrasing my argument this way, it is puzzling how Waldron could think that I do not take the position that an upstream speech restriction cannot “destroy” the legitimacy of downstream legislation. See supra note 8.
. For example, it is morally outrageous that most salaried employees in the United States making more than $15,000 per year pay some income tax, while some billionaires pay none. Nonetheless, I am not prepared to say that these employees have a moral right as opposed to a good moral reason not to be assessed income tax.
. In that Scenario, it will be recalled, Lu significantly curtailed the ability of P (Elaine) to speak out against Ld, whose application infringed Elaine’s fundamental right to religious liberty. This confluence of procedural and substantive moral deprivation gives rise to strong moral reason against enforcement. But as I discussed in my Opening Article, the same-sex couple’s interest in equal treatment in places of public accommodation supplies a strong countervailing moral reason supporting enforcement. The point of the Scenario is that the speech restriction has changed the moral valence of enforcement from positive to negative. Consistent with Waldron’s reformulation of my position, one could conceive the EP Scenario as presenting a clash of moral rights, with the right against enforcement prevailing. With such strong moral claims on both sides, however, I prefer to eschew such rights talk and simply say that, on balance, the restriction on Elaine’s ability to oppose the antidiscrimination measure renders immoral its enforcement against her for refusing to photograph the wedding. A clearer case for a speech restriction giving rise to a moral right against enforcement would occur in the absence of such a strong moral reason supporting enforcement. This might be the situation, for instance, with enforcement of a law forbidding polygamy against someone whose opportunity to speak out against the ban on plural marriage had been severely curtailed.
. Of course, this does not mean that Elaine in fact has a positive law right that Ld not be enforced against her. The existence of such a right depends on particularities of the legal system in which the right is claimed, including the availability of higher-order positive law such as a constitutional provision on which an exemption can be based. In using the term “rights” in his summary of my position, Waldron is obviously referring to moral rights, not positive law rights. As explained in the text, however, his description of my position might erroneously attribute to me the view that there should in all cases in which Lu has rendered enforcement of Ld immoral be a positive law right against such enforcement.
. See Weinstein, supra note 1, at 738 n.92. Waldron also describes my position as holding that P “has a right to disobey” Ld. Waldron, supra note 2, at 709. But I said nothing about a “right” to disobey, only that Lu destroyed any political obligation that P may have had to obey that law, that is, to obey the law just because it is the law. See Weinstein, supra note 1, at 535. Perhaps having been influenced by Waldron’s Response, Professor Eric Heinze also reads me as arguing for a “right of disobedience . . . by a citizen whose views are excluded” from public discourse by a speech ban. Heinze, supra note 3, at 644. Expressly citing Waldron’s Response, Shiffrin also thinks my argument “confers a license” on P to disobey the law. Shiffrin, supra note 3, at 682. Concededly, and analogous to what I said about Elaine having a moral right not to have the antidiscrimination law enforced against her, indeed precisely because this enforcement would be immoral, I think Elaine does have a moral right not to obey the law’s requirement that she photograph the same-sex wedding. But as with the enforcement question, this conclusion does not entail the view that she should have a positive law right to an exemption from the law’s operation. (Cf. Heinze, supra note 3, at 644, attributing to me the view that those who have been excluded from public discourse by a viewpoint-based speech restriction should have “immunity” from, and not “face the legal consequences” for disobeying, the downstream law.) And it obviously does not mean that she in fact has such a right.
. That Lu can damage short of destroying the legitimacy of enforcing Ld against P also shows why I do not, as Shiffrin contends, “give up the ghost” with respect to the effects on legitimacy resulting from restrictions on racist speech. Shiffrin, supra note 3, at 681. Shiffrin notes that I specify that “the problem of justifying coercion to a free and autonomous person” arises only when P can reasonably disagree with Ld. But, he continues, since it would be extremely difficult if not impossible “to formulate a persuasive case that arguments based in racial prejudice amount to reasonable disagreement,” my argument about free speech and legitimacy has no bearing with regard to restrictions on racist speech. Id. To begin with, I think Shiffrin too facilely assumes that while there can be reasonable disagreement about whether it is moral to legally force a person to photograph a same-sex wedding in violation of her religious convictions, there can be no reasonable disagreement about forcing someone to photograph an interracial marriage in violation of her religious dictates. Much would depend on whether the racial discrimination in question constitutes “racial prejudice.” Cf., a photographer who belongs to a Black Nationalist religion refusing to photograph an interracial marriage. But even if Shiffrin is right about this, restrictions on racist speech might nevertheless diminish without destroying the legitimacy of enforcing antidiscrimination laws against a racist whose ability to express his authentic reasons for opposing these laws was curtailed by an upstream speech restriction.
. Thus Waldron writes that he wants to “dispute the whole argument” that upstream speech restrictions can have the “deontic effect” on the “rights” of people regarding downstream legislation that I suppose. Waldron, supra note 2, at 707. He adds that “suppose one were to concede that hate speech laws have a deleterious impact on the quality of the political process . . . [t]he most I would concede is that something has gone wrong with the character of public debate overall,” but denies that the “moral effects” of this speech restriction can generate particular rights. Id. In contrast to his view that speech restrictions cannot affect the normative legitimacy of a particular downstream law in the way I contend, Waldron is willing to go so far as to not deny that speech restrictions can affect systemic legitimacy. Id. at 711.
. Thus as Dean Robert Post aptly observes in his Commentary: “It is not sufficient to observe that members of a legislative assembly are free to express their opposition to the statute . . . . Freedom of speech underwrites democratic legitimation when it allows persons to participate in the formation of public opinion.” Post, supra note 3, at 654. See also id. (“Elections are only an ‘intermittent mechanism,’ whereas public opinion is ‘constantly active’ and, ‘in the long run,’ can exercise ‘a great and growing influence,’” (quoting 3 James Bryce, The American Commonwealth 159 (New York, MacMillan & Co. 1888))).
. Id.. As discussed, supra, text accompanying notes 10 to 14, by problematically describing his disagreement with me as centering on whether upstream speech restrictions can create “rights” in people not to have downstream laws enforced against them or to disobey these laws, Waldron may exaggerate somewhat the extent of this disagreement.
. Id. at 708. In limiting the importance of public discourse to its ability to persuade members of the legislature on how to vote, Waldron’s conception of the purpose of freedom of expression is remarkably, and in my view, unjustifiably narrow. For one, it ignores the interest of a speaker in confirming “his or her standing as a responsible agent in, rather than a passive victim of, collective action.” Dworkin, supra note 9, at vii.
. Accord, Robert C. Post, Citizens Divided 31–42 (2014). Note that Waldron’s minimalistic view of the legitimation of free speech as compared to the legislative process would apply not just to particular downstream laws but to systemic legitimacy as well.
. Waldron’s scenario is inapt in another way. It examines the effect of disenfranchisement on a species of systemic legitimacy, that is, on all the laws passed in a particular legislative session. My concern, in contrast, is with the effect of speech restrictions on particular downstream laws. Waldron curiously claims that I am “not happy with this systemic approach to legitimacy.” Waldron, supra note 2, at 711. Far from being unhappy with such an approach, in a previous article on freedom of expression and political legitimacy I emphasized how viewpoint-based speech restrictions can undermine systemic legitimacy. See James Weinstein, Free Speech and Political Legitimacy: A Response to Ed Baker, 27 Const. Comment. 361 (2011). The reservation that I expressed in my Opening Article was that the damage to systemic legitimacy claimed to be caused by hate speech itself is incommensurable with, or at least very difficult to compare to, the detriment to legitimacy of particular laws that I assert can be caused by hate speech restrictions. See Weinstein, supra note 1, at 577. Waldron’s scenario suffers from a similar problem. An isolated deprivation of the equal opportunity to participate in the political process, be it from a voting or a speech restriction, will have a negligible effect on the entire political system or, as in Waldron’s scenario, even on the entirety of the laws passed during a legislative session. As such, his scenario tells us little about, and in any event is hard to compare to, the effect that a deprivation of the equal opportunity to participate in the political process might have on the legitimacy of a particular law. Cf. infra note 36 and the scenario in text accompanying it.
. Interestingly, it has been alleged that some actual restrictions on early voting or ID-requirements were in fact imposed to disenfranchise people because of their political affiliation or even race. David A. Graham, North Carolina’s Deliberate Disenfranchisement of Black Voters, The Atlantic (July 29, 2016), https://www.theatlantic.com/politics/archive/2016/07/north-carolina-voting-rights-law/493649. Such motivation would, of course, raise greater legitimacy concerns than in Waldron’s scenario, and perhaps this is why he didn’t mention it. Indeed, disenfranchisement on the basis of race can have a ruinous effect on systemic legitimacy. For this reason, I think a good case could be made that the longstanding practice of disenfranchising African-Americans in various jurisdictions in the United States until the latter part of the twentieth century rendered the entire legal system in these jurisdictions illegitimate as to these disenfranchised citizens, making immoral the application to them of all laws which with people could have reasonably disagreed. As a result, such disenfranchisement arguably gave those disenfranchised people a moral right to, in Waldron’s terminology, “rise up in revolution.” Waldron, Political Legitimacy, supra note 4, at 332. Shiffrin suggests that the current American legal system is illegitimate, particularly as regards African-Americans. Shiffrin, supra note 3, at 679-80. He therefore concludes that “[i]f revolution is not justified in the United States, it is not that the government is worthy of our respect” but because of “pacifist principles” or that such a revolution would “be unsuccessful or cause more harm than good.” Id. at 679 n.19. I wholeheartedly agree with Shiffrin that there are currently very serious deficits in the legitimacy of the legal system in the United States, including the existence of government “lobbyists for the rich at the expense of the poor” and “police departments with cultures designed to cover up the police murders of people.” Id. at 679–80. Contrary to Shiffrin, however, I do not believe that the deficits in the current legal system justify revolution even by those most aggrieved. This is in no small part because of the right of every American to express virtually any view in public discourse, including vehement condemnation of the police, as well as the right of each citizen to cast an equally-weighted vote. See Reynolds v. Sims, 377 U.S. 533 (1964).
. Concededly, the hypothetical ban on publicly opposing the tax increase obviously curtails the ability of speakers to express their views on a particular subject more than does even the most restrictive hate speech provision currently in force in any democratic country. It therefore would have far more detrimental effect on the legitimacy of downstream legislation than do hate speech laws. As explained above, I want as much as possible in this Part to put to one side the issue of hate speech bans, as Waldron largely does in the part of his response challenging my basic premise that speech restrictions have the potential to undermine the legitimacy of the enforcement of downstream laws against particular individuals. The hypothetical, then, is not offered as comparable to hate speech bans but rather only to show that if the speech restriction is severe enough, it can indeed have the ruinous effect on downstream legislation that Waldron denies that it can. Whether under certain circumstances hate speech bans can have such an effect is a question that I address in Part II.
. Such voting restrictions are analogous to unreasonably burdensome content-neutral time, place and manner restrictions. Although these restrictions can unduly impair the right to participate in public discourse, see, e.g., United States v. Grace, 461 U.S. 171 (1983) (invalidating law prohibiting demonstrations on sidewalk in front of Supreme Court building), they ordinarily do not raise the equality concerns presented by viewpoint-discriminatory restrictions on free speech.
. Waldron correctly notes that I do not take a position on whether an upstream speech restriction can affect the application of a downstream law to persons who had no desire to speak out against it. Since I wanted in my analysis to focus on the impact that viewpoint-based laws might have on the interests of speakers or would-be speakers to participate in public discourse, I did not consider the impact that such speech restrictions might have on audience interests. For such a discussion, see James Weinstein, Speech Categorization and the Limits of Free Speech Formalism: Lessons from Nike v. Kasky, 54 Case W. Res. L. Rev. 1091, 1117-33 (2004).
. I say “apparently” because at this point in his analysis, Waldron is focusing on the effect of hate speech bans on the interests of racists restrained by such laws. I think though it is a fair inference that Waldron would come to the same conclusion as to any speech restriction that could conceivably be enacted in any contemporary liberal democracy.
. Though it might slightly reduce the “reservoir” of the legal system’s legitimacy. See Weinstein, supra note 33, at 368 n.24 (quoting Robert A. Dahl, Polyarchy: Participation and Opposition 148–49 (1971)). Despite my explicitly making this point in my Opening Article, see Weinstein, supra note 1, at 574 n.165, Heinze seems to read me as contending that viewpoint-based speech restrictions on public discourse can destroy the legitimacy of the entire legal system as to that citizen. “For Weinstein—although this is not exactly his phrasing—a democracy that excludes citizens from democratic opinion formation effectively dissolves its social contract with them, relieving them of their duty to obey law.” Heinze, supra note 3, at 638. See also supra note 9 (discussing Heinze’s view that Dworkin was concerned with the effect of hate speech restrictions on systemic legitimacy). For there to be any such catastrophic consequence, the democratic deprivation would have to be far greater in scope and impact than any speech restriction currently in force or likely to be enacted in any contemporary mature and stable democracy. See supra note 34 (arguing that the longstanding practice of disenfranchising African-Americans in various jurisdictions in the United States until the latter part of the twentieth century may have rendered the entire legal system in these jurisdictions illegitimate as to these disenfranchised citizens).
. Waldron, supra note 2, at 711. In my Opening Article, I explained that “[i]ndividuals . . . are entitled to participate as political equals not just to vindicate their personal interests narrowly defined, but also in deciding what in their judgment is best for society as a whole.” Weinstein, supra note 1, at 536.
. Heinze argues that any attempt to limit my argument to downstream laws “‘directly’ related to the proscribed speech, such as laws imposing upon employers various non-discrimination norms contrary to the viewpoints of excluded hate speakers” is “too arbitrary.” Heinze, supra note 3, at 644. In his view, once a citizen is excluded “pro tanto from public discourse, there is no area of law to which that exclusion becomes irrelevant in principle.” He gives an example of an anti-Semite who believes that “Jews run the world” and who thus can connect that belief to “any legal norm or practice,” including the speech limit or the tax on cigarettes. I am not sure I see the problem. If an anti-Semite were prohibited from giving as his reason for opposing certain laws his authentically held belief they were masterminded by a Jewish conspiracy for the benefit of Jews, then such a speech restriction would most likely damage the descriptive legitimacy of those laws as applied to this bigot and in certain situations impair, and perhaps even on a very rare occasion destroy, normative legitimacy as well. This potential damage would usually be less if a bigot did not want to speak out against a particular law or laws but was nonetheless prohibited from expressing the view that all laws and policies in his jurisdiction are unduly influenced by a group of people. And of course, many of those wanting to engage in hate speech, or expression that many believe to be hate speech such as proclaiming that homosexual conduct is immoral, do not have such “a comprehensively conspiratorial world view.” Id.
. Weinstein, supra note 1, at 582. I added that if there were non-coercive measures that had not yet been tried to combat the alienating effect of hate speech that Waldron described, then this argument “weighs heavily” against the propriety of hate speech laws in a free and democratic society. Id at 583.
. I am referring here to moral as opposed to legal justification. Even an immoral law might as a matter of positive law be legally justified in the sense, for instance, that it comports with constitutional constraints. See infra note 62.
. Consistent with this view, in my Opening Article I concluded that despite their viewpoint-discriminatory effect, hate speech laws limited to prohibiting use of highly vituperative language such as referring to members of minority groups as “rats” or cockroaches” or other animals “‘we would normally seek to exterminate’” to “stir up racial hatred” might in some circumstance be “adequately justified.” Weinstein, supra note 1, at 547–48. Contrary to Waldron, however, I contended such laws nonetheless had some non-trivial effect on normative legitimacy, though not enough to “substantially diminish” anyone’s political obligation to obey a downstream antidiscrimination law and “not nearly substantial enough to nullify the large moral benefit” of preventing landlords from refusing housing to people based or their race or ethnicity. Id. at 548. Importantly, this assessment was made on the assumption that these laws would in actual operation precisely target just such highly vituperative speech. Accordingly, I assumed for the sake of argument that these laws would not substantially impair the ability of individuals without excessive vituperation to express opposition to antidiscrimination measures or to criticize the people these measures are meant to protect. Consistent with what I say below, if experience showed that these laws were in actual operation applied in a way that substantially interfered with such expression to the detriment of political legitimacy of these downstream laws, then the justification of the upstream speech restriction would need to be reevaluated.
. Dennis v. United States, 341 U.S. 494 (1951). Of course, the Court’s decision to uphold the law and affirm these convictions means only that a majority of the Justices thought the law was constitutionally justified. But there can be little doubt that several of the Justices also thought that the law was morally justified.
. Jeremy Waldron, Free Speech & the Menace of Hysteria, N.Y. Rev. Books, May 29, 2008, http://www.nybooks.com/articles/2008/05/29/free-speech-the-menace-of-hysteria/ (reviewing Anthony Lewis, Freedom for the Thought That We Hate: A Biography of the First Amendment (2008)).
. Interestingly, Waldron seems to concede that Britain’s hate speech laws, which he obviously thinks are very well justified, could have a “minimal effect on legitimacy.” Waldron, Political Legitimacy, supra note 4, at 335. So wouldn’t it follow that a just-barely justified law could have a greater impact on legitimacy? (In fairness to Waldron, he also has characterized the loss of legitimacy caused by Britain’s hate speech laws as “minimal or nonexistent.” Waldron, Hate Speech, supra note 4, at 183 (emphasis added)). In any event, Waldron previously seems to have thought the question of the relationship between justification of a speech restriction and its effect on the legitimacy of downstream laws was not quite as clear cut as he now claims it to be.
. By way of analogy, if a bombing campaign as part of a just war is morally justified, then so is all collateral damage that could not have been avoided with the exercise of due care. On this view, while the collateral damage is regrettable, it is not immoral.
. So if in the bombing campaign scenario posited in note 68 supra it was determined after several air raids both that the civilian casualties were far greater than anticipated and that the strategic value of the target was less than originally thought, continued raids might well be immoral at least in the absence of increased precautions against collateral damage.
. By the same token, those of us who are skeptical that hate speech laws are justifiable in mature, stable democracies should be open to revising views in light of new empirical evidence as well as new arguments, or better argued versions of old ones. While I remain skeptical about the operation in actual practice of even hate speech laws limited to highly vituperative racist expression, Waldron’s arguments have persuaded me that as a theoretical matter such bans might be justified. And as I make clear in Part II of this Reply, if there were persuasive evidence that hate speech in public discourse was actually deterring a significant number of members of some vulnerable minority group from participating in public discourse, I would be open to the propriety of even broader hate speech laws after all reasonably available alternatives to speech suppressive measures had been tried and found wanting.
. Heinze thus begins his Commentary by stating that “[d]emocracy is the ongoing product of public discussion” and then asking “[w]ith what legitimacy, then, can a democracy limit its citizens’ participation in that discussion?” Heinze, supra note 3, at 631. See also id. at 633, 636, 639, 643, 645, and 649.
. Id. at 664. Schauer is not sure how “translating this basic proposition of normative political philosophy into the language of ‘legitimacy’ adds very much” but acknowledges that this objection is basically a “terminological quibble.” Id.
. See Weinstein, supra note 1, at 537 n.42 (“studies find that people’s increased belief in their having an obligation to obey the law results in their voluntary compliance with the law.”). Even my slightly more nuanced statement that “the weakening of this sense of obligation would likely lead to less compliance with the law,” id. at 547 n.82, overstates for the reason Schauer gives what can fairly be inferred from the empirical evidence.
. Post, supra note 3. Indeed, because he considers the question of normative legitimacy to be properly within the realm of moral philosophers, not legal scholars, Post eschews this inquiry altogether. To the extent that Post means to imply that it is not properly within the scope of a legal scholar’s bailiwick to consider how normative legitimacy might be affected by speech restrictions, I respectfully disagree. Free speech issues are suffused with normativity. Though always of some relevance, the deep moral underpinnings of free speech sometimes properly remain submerged, such as with highly technical inquiries about doctrinal fit. At other times, however, consideration of the normative core becomes crucial, such as in attempting to explain, as I try to do in this Symposium, why democratic self-governance is a vital free speech value. As a preeminent legal scholar has written: “The normative essence of democracy is thus located in the communicative processes necessary to instill a sense of self-determination.” Robert C. Post, Racist Speech, Democracy, and the First Amendment, 32 Wm. & Mary L. Rev. 267, 282 (1991). Following Post’s lead, I endeavor in this Symposium to elucidate this essence.
. See Weinstein, supra note 1, at 541 (“It is the thesis of this article that the infringement of this fundamental interest of equal political participation can have severe consequences not just for the legitimacy of the legal system but also for individual downstream laws.”).
. I noted that other types of speech restrictions, such as blasphemy laws and bans on glorifying terrorism or aiding terrorist organizations, also have the potential to undercut the legitimacy of downstream laws. Id. at 541 n.60.
. For such a discussion, see James Weinstein, Hate Speech, Pornography and the Radical Attack on Free Speech Doctrine (1999). As noted above, the only overall implication for the hate speech debate that I drew from my analysis is that the potential damage to the legitimacy of downstream legislation “weighs against” the propriety of hate speech laws in mature and stable democracies, and “heavily” against such laws before non-coercive remedies have been exhausted. See supra note 51 and accompanying text.
. See Waldron, supra note 2, at 704. Far from finding hate speech laws wrong in principle, I wrote in my Opening Article that a legislature could under some circumstances reasonably conclude that narrow bans on highly vituperative hate speech of the type that Waldron supports are justified. See Weinstein, supra note 1, at 548. My doubts about such restrictions are not so much theoretical as they are pragmatic. So Shiffrin is correct in observing that I “might be counted out as a general opponent of hate speech restrictions.” Shiffrin, supra note 3, at 681. Cf. Heinze, supra note 3, at 635 (arguing that all viewpoint-based restrictions on speech, including hate speech bans, are contrary to basic democratic principles). Regarding my position on hate speech laws, Brown writes the impression he gets from my Opening Article is that I “partly” see defenders of hate speech bans as “defend[ing] the indefensible.” I am not sure what to make of the oxymoronic qualification “partly.” (Cf. “I partly think that people who commit hate crimes are deplorable.”) But while I find the breadth of the bans on hate speech that Brown endorses difficult to square with core democratic principles (see infra note 186), I don’t find even these restrictions “indefensible” in light of the serious harm that hate speech is capable of causing even in mature, stable democracies. What I do find indefensible are certain rationales that have been offered in support of hate speech bans, including Brown’s use of the precautionary principle to support the “silencing argument.” See Weinstein, supra note 1, at 757. (I reply in detail below to Brown’s further attempt to defend his use of the precautionary principle.) For another indefensible rationale for suppressing hate speech, see R. v. Keegstra,  3 SCR 697, 758, in which writing for a majority of the Canadian Supreme Court Justice Robert George Brian Dickson stated: “Hate propaganda seriously threatens . . . the enthusiasm with which the value of equality is accepted and acted upon by society.”
. Apropos the difficulty of this task, Gelber aptly adds “to be fair” that “the lack of a clear definition of hate speech is in fact part of the problem” that I have with hate speech laws in actual operation. Id.
. As discussed below, several participants validly took issue with my using prosecution of anti-homosexual speech under Section 5 of the UK’s Public Order Act, which as I made clear is not a hate speech law, to impugn the operation of hate speech laws. This problem, however, did not derive from lack of a “clear conception” of hate speech. Similarly, the disagreement between Waldron and me, also discussed below, about whether the section of the Public Order Act enhancing the penalty for racially or religiously motivated violations of Section 5 is a hate crime rather than a hate speech statute, does not stem from uncertainty as to what type of expression constitutes hate speech.
. Gelber is mistaken in asserting that I imply “quite strongly” that “two of the defining features of hate speech, properly understood, are vituperation and the use of epithets.” Gelber, supra note 3, at 620. Because Waldron focused on these features in his work, I did so as well, not because I thought that these were “defining features of hate speech, properly understood” but to engage Waldron on his own terms. It is true that I concluded that if highly vituperative hate speech could be narrowly targeted with no substantial chilling effect on those who wanted to express bigoted ideas without vituperation, such a restriction would have a minimal impact on normative legitimacy. Weinstein, supra note 1, at 545. But as important as the impact on normative legitimacy is to the calculus on the propriety of banning hate speech, it is still only one of many factors, including consideration of the impact on descriptive legitimacy which I concluded could be substantially impaired by even the narrow bans that Waldron supports.
. For what it’s worth, I agree with my fellow skeptic Robert Post that whether such restrictions are appropriate depends on “variables like the number of persons in target groups, the intensity of their sense of exclusion” and whether or not “the ambient legal and social environment makes members of target groups feel safe and included.” Post, supra note 3, at 657–58.
. Waldron, supra note 2, at 699–700. See also id. at 706 (suggesting that my “theory of political legitimacy” is “specifically invented for” the hate speech debate and “rigged” to yield a certain result.)
. See, e.g., James Weinstein, Database Protection and the First Amendment, 28 U. Dayton L. Rev. 305, 317, 327 (2003); Weinstein, supra note 40 at 1104, 1113-15; James Weinstein, Fools, Knaves and the Protection of Commercial Speech: A Response to Professor Redish, 41 Loy. L.A. L. Rev 133, 145 (2007); James Weinstein, Institutional Review Boards and the Constitution, 101 Nw. U. L. Rev. 512–13 (2007); James Weinstein, Participatory Democracy and Free Speech, 97 Va. L. Rev. 491, 498, 509-14 (2011).
. See Weinstein, supra note 33. But not only is Waldron’s motivational concern misplaced, it is also odd. Why on earth would any right thinking person want to “rig up” an argument to protect something as noxious as racist, homophobic, or anti-Semitic expression? To the contrary, the natural temptation would be to “rig” the argument against the protection of such loathsome expression as evinced, for example, by Alexander Brown’s invocation of the precautionary principle. See Part II. C. 2, infra. I would hope that Waldron and other defenders of hate speech legislation would understand that the motivation for my skepticism of hate speech bans is the concern that these restrictions and the justifications offered to support them will undermine the strength of a free speech principle required to adequately protect dissent in a free and democratic society, including by members of minority groups. See Part II. C. 2, infra. In this regard, it is worth noting that Mari Matsuda, a prominent supporter of hate speech bans, decades ago expressed admiration for “the conviction and conviction [of] Jewish civil libertarians who have eloquently, and at great personal cost, argued for the free speech rights of Nazis and Klan members.” Although she passionately disagreed with this position, she recognized that it did not derive from insensitivity to the harms of hate speech but rather from the belief that “the right of protest [is] essential for the protection of minorities.” Mari Matsuda, Public Response to Racist Speech: Considering the Victim’s Story, 87 Mich. L. Rev. 2320, 2326 (1989).
. Also giving rise to this worry about “rigging” is Waldron’s erroneous view that in declining to balance the loss of legitimacy to a particular law against gains to systemic legitimacy as argued for by Alexander Brown due to incommensurability problems, I was somehow “dismiss[ing] or ignor[ing]” Brown’s concern and therefore not following “the legitimacy principle where it leads.” Waldron, supra note 2, at 699–700. Far from dismissing or ignoring Brown’s concern, I dealt with it at considerable length. See Weinstein, supra note 1, at 576–78. In addition, to obviate the incommensurability problem I then built on Brown’s argument to identify a commensurable countervailing legitimacy concern. Id. at 580–81. Indeed, I concluded this countervailing legitimacy concern arising from hate speech itself might, like the one I have identified arising from a hate speech ban, also destroy the legitimacy of a downstream law as applied to members of this minority group. Id. at 581. Later in this Reply, I will further engage Brown’s countervailing legitimacy concerns. See infra text accompanying notes 169 to 203. In addition, now that the discussion has been expanded to also focus on hate speech and systemic legitimacy, I will try to evaluate the claims that commentators have made about the effect of both hate speech bans and hate speech itself on the legitimacy of the legal system. See notes 236 to 245, infra. For unlike the determent of hate speech bans on particular laws versus the damage of hate speech itself to systemic legitimacy, this expanded discussion presents commensurable claims. In short, whatever problems there may be with my view about free speech and legitimacy, I don’t think that failing to follow “the legitimacy principle where it leads” is one of them.
. Shiffrin writes that “a restriction on hate speech is not imposed merely because government finds the speech disagreeable, disturbing, or offensive as Weinstein suggests.” Shiffrin, supra note 3, at 677 n.68. This statement is misleading. What I actually wrote, quoting from an earlier article not focusing on hate speech, is that “if an individual is excluded from participating in public discourse because the government disagrees with the speaker’s views or because it finds the ideas expressed too disturbing or offensive, any decision taken as a result of that discussion would, as to such an excluded citizen, lack legitimacy.” Weinstein, supra note 1, at 529–30 (quoting Weinstein, Participatory Democracy and Free Speech, supra note 111, at 498). Neither in my Opening Article nor in the article it quotes did I state that hate speech restrictions are imposed “merely” for these reasons. To the contrary, in the quoted article I write that hate speech “undoubtedly causes” psychic injury to members of minority groups and noted the possibility that such expression might lead others to discriminate against minorities. See Weinstein, supra note 1, at 500. What I did claim in my Opening Article is that indefensible reasons offered in support of hate speech bans, such as Brown’s invocation of the precautionary principle, raise the suspicion that the real motivation for suppressing such speech is the “morally repugnant” viewpoint it expresses. Weinstein, supra note 1, at 580. See also id. at 579 n.183.
. Waldron is correct that proponents of the legitimacy argument, including me, have sometimes been “loose” in describing the nature and extent of curtailment on the ability of people to participate in public discourse imposed by speech restrictions. Waldron, supra note 2, at 700. He notes, for instance, that in my Opening Article I quote an earlier work in which I referred to people being “excluded” from participating in public discourse. Id. But far from repeating this language to be as “loose” as I think I can “get away with on this matter,” id., I quoted this earlier, broad formulation at the beginning of that article to “credit” Waldron for “properly criticizing Dworkin and me for not adequately specifying” our claim. See Weinstein, supra note 1, at 531. So Waldron should learn to take “yes” (and a compliment) for an answer.
. Waldron, supra note 2, at 701 (citing Waldron, Hate Speech supra note 4, at 150-55). Viewpoint-based speech restrictions are a subset of content-based laws. See Rosenberg v. Rector & Visitors of Univ. of Va., 515 U.S. 819, 829 (1995) (“Viewpoint discrimination is . . . an egregious form of content discrimination.”). Thus, in addition to including viewpoint-discriminatory laws, the category of content-based speech restriction comprises laws that curtail expression based on its subject matter, its use of particular words or symbols or, more generally, because of the communicative impact of the speech.
. While a requirement that the speech must intend to create a certain effect cannot eliminate the content-based nature of a speech restriction, it is relevant to the extent of the curtailment it imposes on a speaker’s ability to express a particular idea in public discourse. This is the issue explored in Part II C, below. In this regard, however, it should be noted that it is not at all clear that the British hate speech law, let alone most hate speech provisions, imposes a requirement that the speech intend to have the effect of stirring up hatred against a certain group of people. See Post, supra note 3, at 653.
. Waldron, Hate Speech, supra note 4, at 151. Waldron’s argument that hate speech laws are not content based is similar to the view, noted but “put aside” by Stone, that hate speech laws are not viewpoint based because “they are not targeted at viewpoints per se but at the harmful consequences of hateful speech.” Stone, supra note 3, at 689 n.7. Shiffrin may be making something akin to this argument in noting that although the United States Supreme Court would find hate speech bans to be “impermissible point-of-view discrimination,” the better approach “would be to recognize that racist speech causes unjustifiable harm and promoting racial tolerance does not.” (Perhaps though Shiffrin is not denying that hate speech laws are viewpoint discriminatory but only that they are not impermissibly so.) For all practical purposes, this position renders the category of viewpoint discriminatory laws the null set. I take it that every participant in this Symposium agrees that it is not acceptable in a liberal democracy for government to suppress speech in public discourse because it “disagrees with a speaker’s views or because it finds the ideas expressed too disturbing or offensive.” See supra note 115 and accompanying text. Accordingly, all restrictions on public discourse need to be justified in terms of “the harmful consequences” of the speech. But if the harm in question is claimed to result from the expression of a particular point of view—be it that democratic governments should be violently overthrown and replaced with the dictatorship of the proletariat; that those drafted to fight an unjust war should refuse to serve; or that certain racial, ethnic and religious groups are inferior—then no matter how real and substantial the harm in question, a law suppressing that point of view nonetheless is still based on that viewpoint.
. Waldron, supra note 2, at 701–02. Post and Heinze in their Commentaries vigorously object to the distinction between the manner and content of expression on which Waldron relies. Invoking Percy Bysshe Shelley’s argument about the impossibility of translating poetry, Post argues there are “ideas which can be expressed only in the particular outrageous style that hate-speech regulations proscribe,” Post, supra note 3, at 656. Heinze similarly derides what he calls the “form-content” distinction, arguing at length that the suppression of particular words, including racial epithets, inevitably suppresses ideas. Heinze, supra note 3, at 649–50. I agree that the regulation of the manner of expression, including bans on vituperation, can sometimes substantially impede speakers from expressing the precise idea they want to convey. For example, there is no other way to express the precise idea conveyed by the message “Fuck the Draft” emblazoned on an anti-war protestor’s jacket. Still, I do not agree that the regulation of the manner of expression has an inevitable suppressive effect on a speaker’s ability to convey ideas as Post and Heinze contend. In his Commentary, Blasi attacks the form-content distinction from another direction, persuasively arguing that temperate hate speech poses a greater danger to the civic standing of vulnerable minority groups than does vituperative expression. Blasi, supra note 3, at 589. Accord, Weinstein, supra note 103, at 130 (conjecturing that “subtly racist public expression” is more likely to instill racist beliefs in others than are pronouncements by “gruesome characters as neo-Nazis, skinheads and Klansmen.”
. I acknowledged in my Opening Article that despite their discriminatory effect, hate speech laws banning only highly vituperative expression might arguably be considered viewpoint neutral. Weinstein, supra note 1, at 545 n.79. However, I rejected that characterization because such laws would likely have a discriminatory effect in that such bans would, for instance, prevent a bigot agitating against an antidiscrimination measure from referring to members of minority groups as “cockroaches” but would not prevent supporters of the antidiscrimination measure from using such epithets to refer to bigots. Id. at 545. Nonetheless, I concluded that because a legislature could reasonably assume that referring to members of vulnerable racial and ethnic groups as animals we normally seek to exterminate could make racial and ethnic minority groups unsure of their status in society, the discriminatory effect might under some circumstances be justified. Id. at 548. I therefore find it bewildering that Gelber could write that my “steeping in the requirement under First Amendment jurisprudence to avoid viewpoint discrimination at all costs blinds [me] to the differential harms of these two events.” Gelber, supra note 3, at 626.
. In contrast, even if limited to bigoted views expressed with extreme vituperation, under American free speech doctrine such a ban would most likely be considered viewpoint based. See R.A.V. v. City of St. Paul, 505 U.S. 377, 391–92 (1992). In his Commentary, Shiffrin points to speech regulations in the United States that he claims to be viewpoint discriminatory but which do not offend the First Amendment. Shiffrin, supra note 3, at 676. With the exception of certain justifications that have been offered to justify obscenity laws, which I agree are arguably viewpoint discriminatory in the relevant sense (see generally James Weinstein, Democracy, Sex and the First Amendment, 31 N.Y.U. Rev. L. & Soc. Change 865, 893-96 (2007)), the other two examples he gives are not analogous to the suppression of hateful ideas about racial, ethnic or religious groups. Shiffrin notes that under the defamation law “nasty” comments about a person are sometimes actionable, while “nice” statements are protected by the First Amendment, and that similarly, under some circumstances advocacy of illegal action is unprotected speech, while advocacy of legal action is protected. Shiffrin, supra note 3, at 677. While these distinctions may in some sense be based on a speaker’s viewpoint, they do not single out a particular ideological position for suppression as do hate speech laws. As such, they do not have the same implications for political legitimacy as do hate speech bans.
. The one important exception is if the selective imposition of the manner restriction is not well justified. Suppose, for instance, that anti-war protestors were not allowed to shout profanities such as “Fuck the War” in public, while those who supported a country’s war effort could yell “Fuck the Enemy.” In that event, there would be a significant effect on normative legitimacy even though the anti-war protestors could still convey the basic propositional content of their message. For this reason a poorly justified selective regulation of the manner of expression would be viewpoint discriminatory as I use that term in this discussion. However, since I believe that a selective ban on highly vituperative hate speech, such as referring to members of certain vulnerable ethnic and racial groups as “animals we normally seek to exterminate,” Waldron, Political Legitimacy, supra note 4, at 335, might in certain circumstances be well justified, there would in that case be no significant detriment to normative legitimacy worked by the selectivity. Still, as I emphasize in my Opening Article, even if such well-justified restrictions do not significantly damage normative legitimacy, they still might have a significant detrimental effect on descriptive legitimacy. See Weinstein, supra note 1, at 564–65.
. Waldron seems to have upon reflection also slightly modified his position about restrictions on the manner of expression and normative legitimacy. He acknowledges that “[I]f a citizen thinks of himself as the sort of person who shouts ‘fuck’ or utters threats in political debate or shows dirty pictures during his political orations then—I don’t know—maybe a case can be made that he is not being respected as such.” Waldron, supra note 2, at 713. (He adds that he is “being respected as someone who could be better than that, and as someone who has responsibilities as well as rights in the political process,” but realizes that this is “another matter.” Id.) So if someone wanting to express a particular viewpoint, say an anti-war or anti-abortion protestor or someone wanting to express racist views, is selectively forbidden from using profanity in a public debate, one might infer that Waldron now agrees with me that a hate ban limited to vituperation has somewhat greater impact on normative legitimacy than he previously thought. Shiffrin, in contrast, does not even make even this minimal concession, claiming that hate speech restrictions do not disrespect the speaker or his equal moral worth but at most only show “disrespect for a particular speech choice the citizen would like to make.” Shiffrin, supra note 3, at 687.
. Stone finds it “especially eye-catching” that I invoke the concept of viewpoint discrimination as a core principle of freedom of speech “given that in its aversion to viewpoint discrimination (like much else) First Amendment law is highly unusual.” Stone, supra note 3, at 687. But this aversion to viewpoint discrimination is not just some quaint and peculiar detail of American free speech doctrine. It is, rather, a core doctrinal rule implemented to vindicate the basic democratic commitment to formal equality and the equal moral worth of each citizen and to promote political legitimacy. See Weinstein, supra note 1, at 536–37. While the details of free speech doctrine will inevitably and properly differ in different democratic societies, there are nonetheless core precepts that apply in any democracy, including the right to formal equal participation in the democratic process. I am not claiming that any democracy that fails to share American free speech doctrine’s broad, fierce version of viewpoint discrimination is betraying this core democratic precept. Thus Shiffrin, supra note 3, at 679 n.17, over-reads my position when he claims that I believe that Britain’s hate speech ban necessarily involves “impermissible” viewpoint discrimination. Cultural differences properly have some bearing in determining whether a regulation breaches a core democratic precept in a particular society. Nonetheless, I do believe that viewpoint-discriminatory restrictions on public discourse as I use the term in this article (see supra text accompanying note 129) at least directly implicate this core democratic precept even if they do not violate it.
. And, of course, if the thirty-two cases I cite are empirically insufficient to show that most hate speech laws fail to create such “safe havens,” then Waldron’s assertion that most such laws allow ample opportunity for speakers to express bigoted views without vituperation, a claim he supports by reference to two statutes, is far more empirically deficient. While this should go without saying, I mention it only because Gelber, while properly critical of the empirical limitations of my challenge to Waldron’s claim, gives Waldron a pass on the far greater empirical shortcomings of his claim.
. Gelber notes “in particular the cases in which Christians put forward their views about homosexuality.” Id. at 629. Gelber gives no reason for finding these cases particularly problematic as compared to the other ones I discuss. It would be interesting to know why Gelber does not regard as equally problematic, for instance, the fining of an Austrian academic for saying in a seminar that Mohammed had “a thing for little girls.” See supra, text accompanying note 137. It would also be worth knowing whether the call by the Dutch politician to remove “all Surinamers, Turks and other so-called guest-workers from the Netherlands,” see supra text accompanying note 136, is in her view hate speech that may properly be prohibited in a free and democratic society.
. Waldron, supra note 2, at 703. I tend to agree with Waldron about the British hate speech laws. He did not, however, limit his claim to these laws but rather referred to “most” hate speech laws.
. Id. Waldron adds that “opponents of hate speech regulation ought to consider the best case that can be made for regulation of this sort.” Id. at 703–04. I agree with this sentiment and tried to do just that in my one comprehensive discussion of the pros and cons of hate speech laws. See Weinstein, supra note 103. But just as opponents of hate speech legislation should consider the “best case” for such laws, proponents of such laws should consider the “worst cases” of the application and misapplication of hate speech legislation.
. Blasi, supra note 3, at 585. While Waldron is remarkably nonchalant about these cases, Brown pretends the most troubling among them simply don’t exist. Brown begins this exercise in denial by inaccurately claiming that the “vast majority” of the cases I discuss involve not the group defamation laws (sensu stricto) and incitement to hatred laws that he defends but rather “expression-oriented hate crimes,” specifically, public order offences, which he does not. Brown, supra note 3, at 606. In fact, I discuss ten public order act cases, hardly a “vast majority” of the thirty-two cases I mention. Of the remaining twenty-two cases, fifteen involved incitement to hatred laws and seven involved group defamation laws (albeit some not sensu stricto). Brown acknowledges only one of these twenty-two cases, Glimmerveen & Hagenbeek v. Netherlands. He ignores the other twenty-one cases, including the cases against a Catholic Bishop, an Austrian academic, and Brigitte Bardot discussed supra, text accompanying notes 137 to 139.
. See Leslie Kendrick, Speech, Intent, and the Chilling Effect, 54 Wm. & Mary L. Rev. 1633, 1681 (2013) (“[T]he existence of a chilling effect . . . is very difficult to establish, even with the aid of a variety of sophisticated empirical tools.”). Gelber claims that her own research shows “no evidence of chilling effect” from “25 years of the operation of hate speech laws in Australia.” Katharine Gelber & Luke McNamara, The Effects of Civil Hate Speech Laws: Lessons from Australia, 49 Law & Soc’y Rev. 631, 631, 656 (2015). Gelber states that her “analysis of letters to the editor revealed little evidence that public discourse has been diminished over the past 25 years” because “[r]obust debates have been had on a broad range of issues” and that her “analysis revealed the continued expression of prejudice over time.” Id. at 656. She noted that she “detected a shift away from more intemperate styles of language,” but concluded that the shift “cannot be said to support the chilling effect claim.” Id. While it may be true that Gelber’s study shows “no evidence of chilling effect,” it also provides next to “no evidence” that the chilling effect does not exist. That in the aggregate the robustness of discourse in letters to the editor may not have been diminished since the enactment of Australia’s hate speech laws tells us very little about whether particular individuals were deterred by these laws from expressing views not within the intended scope of the prohibition of these laws in these letters or in other settings for public discourse. So to paraphrase Kendrick, Gelber’s study shows that it is “very difficult . . . even with the aid of a variety of sophisticated tools” to establish the non-existence of the chilling effect.
. Cf. Waldron, supra note 2, at 702 (“Weinstein cites the invocation of section 5 of the Public Order Act in a number of British cases . . . to illustrate his contention that hate speech laws make it quite difficult to safely express the ‘basic propositional content’ of bigoted views even when expressed without vituperation or use of vicious epithets.”).
. See Hammond v. Dep’t of Pub. Pros’ns,  EWHC 69, ¶ 5 (Admin), http://www.bailii.org/ew/cases/EWHC/Admin/2004/69.html. It is true, as Waldron says, that analogous to bans on “fighting words,” Section 5 is concerned with public order and not as are hate speech laws, with the “proliferation of racial hatred in a community.” Waldron, supra note 2, at 702. As this and other cases reveal, however, he is wide of the mark in suggesting that as applied Section 5 is not concerned with the proliferation of hatred, or at least intolerance, towards homosexuals and religious groups.
. Norwood v. Dir of Pub. Pros’ns,  EWHC 1564 (Admin), http://www.bailii.org/ew/cases/EWHC/Admin/2003/1564.html. Waldron is correct that this provision is a hate crime law, Waldron, supra note 2, at 703, and for that reason I was careful not to refer to it as a hate speech provision. But in this case, it was applied to speech made criminal by Section 5 and it thus operates with respect to the enhancement as a regulation of hate speech. Waldron mentions that he is “not sure” of my views on hate crime legislation. As I have written, laws that enhance the punishment for bias-motivated crimes do not on their face offend free speech principles but can do so in their application. See James Weinstein, Hate Crime and Punishment: A Comment on Wisconsin v. Mitchell, 73 Or. L. Rev. 345 (1994); James Weinstein, First Amendment Challenges to Hate Crime Legislation: Where’s the Speech?, 11 Crim. Just. Ethics 6 (1992). I should add by way of disclosure that I was an advisor to the State of Wisconsin in its successful defense of hate crime laws in Wisconsin v. Mitchell, 508 U.S. 47 (1993).
. I recognize that Waldron did not claim that any such safe haven actually existed, but only that, what he believes is typical of hate speech laws in other jurisdictions. Britain’s hate speech bans did not extend to non-vituperative expression. I do, however, think it relevant that even if Britain’s hate speech law does not seek to prohibit non-vituperative bigoted expression, other laws do so precisely because the expression is “perceived to be hate speech.” See Gelber, supra note 3, at 627.
. Compare, e.g., Hammond with the prosecution of the Catholic Bishop for calling homosexuality disordered (discussed supra text accompanying note 139) and Norwood with J. Glimmerveen & Hagenbeek v. Netherlands, App. Nos. 8348/78 and 8406/78,  23 Y.B. Eur. Conv. on H.R. 366 (Eur. Comm’n. on H.R.), http://www.bailii.org/eu/cases/ECHR/1979/8.html (discussed supra text accompanying note 136).
. Indeed, in accord with American free speech doctrine’s intense hostility to viewpoint discrimination, a law that selectively punishes hate speech constituting a public order violation is unconstitutional. See R.A.V. v. City of St. Paul, 505 U.S. 377 (1992) (striking down municipal ordinance for selectively targeting fighting words that arouse “anger, alarm or resentment in others on the basis of race, color, creed, religion or gender”). It is sometimes mistakenly (and carelessly) asserted that R.A.V. held that expressive activity at issue in that case—a white juvenile’s placing a burning cross on a black family’s lawn—was constitutionally protected speech. See, e.g., Judith Butler, Excitable Speech: A Politics of the Performative 55 (2013). Not only did the Court not so hold, but it suggested that this “reprehensible” act (R.A.V., 505 U.S. at 396) could have been punished consistent with the First Amendment, under any number of criminal provisions, including a prohibition of terroristic threats. Id. at 380 n.1.
. For this reason, I took issue in my Opening Article to Heinze’s claim such bans can “never promote the state’s democracy.” Weinstein, supra note 1, at 581 n.190. Although not directly responding to this criticism in his Commentary, Heinze continues to insist that viewpoint-based restrictions in public discourse are democratically illegitimate in principle. He thus analogizes such restrictions to a single falsified ballot case in an election which although its impact will almost always be nil, is still contrary to democracy. Heinze, supra note 3, at 645. The analogy is inapt. Unlike a falsified ballot, which can never promote democracy, not even theoretically, hate speech bans can promote democracy if necessary to prevent bigoted expression from, for instance, preventing people from participating in public discourse because they reasonably fear they will be physically harmed if they do so.
. It is true that the line dividing public discourse from other forms of speech may be blurry. Still, certain types of speech are clearly not within the domain of public discourse, including a “Whites Only” sign on a business establishment, see Brown, supra note 170, at 87; a foreman using the word “nigger” to refer to African-American workers, id. at 85; burning a cross on a black family’s lawn, Butler, supra note 167, at 55; or scrawling anti-Muslim graffiti or placing a racist poster on a mosque, Waldron, Hate Speech, supra note 4, at 1–2. Even under American free speech doctrine, none of this speech is protected, and properly so. Accordingly, the harms caused by such expression are irrelevant to whether hate speech that is part of public discourse can appropriately be prohibited under a “silencing” or any other rationale. In contrast, someone standing in a public square with a sign reading “Stop Homosexuality, Stop Lesbianism, Stop Immorality, Jesus is Lord,” see Weinstein, supra note 1, at 555, plainly is engaging in public discourse; as was Brigitte Bardot’s statement on her webpage about Muslims or the Catholic Bishop’s comments to a journalist about homosexuals. See supra, text accompanying notes 138 to 139.
. Brown, supra note 3, at 612. Nearly 20 years ago, I observed that “the proponents of the silencing argument offer no supporting evidence.” Weinstein, supra note 103, at 133. Two years later I wrote that although one could imagine a society in which “a group is so subordinated and racist speech so prevalent that vicious epithets in public discourse will likely impede their participation in democratic self-governance,” there was no evidence that I was aware of that hate speech had this effect in the United States. See James Weinstein, Hate Speech, Viewpoint Neutrality, and the American Concept of Democracy, in The Boundaries of Freedom of Expression and Order in a Democratic Society (T. Hensley ed., 2001). In her Commentary, Gelber concedes that “there is not a great deal” of evidence supporting the silencing argument but notes that “there are findings from psychology that show that individuals subjected to non-physical discrimination suffer significant harms to their physical and mental health.” Gelber, supra note 3, at 623. None of these studies, however, focuses on speech as part of public discourse and therefore provide no support for the claim that such expression causes these harms. In addition, it does not follow that the injuries described in these studies will prevent those subject to “non-physical discrimination” from participating in public discourse. In addition, Gelber writes that her own empirical research “has shown that target communities claim to experience the harms of hate speech alleged in the literature,” citing without further discussion Katharine Gelber & Luke J. McNamara, Evidencing the Harms of Hate Speech, 22 Soc. Identities 324 (2016). But like the other studies she cites, this one also does not distinguish between the harms, including silencing, resulting from hate speech as part of public discourse and non-public discourse such as face-to-face encounters. The only empirical support that Brown offers for the silencing effect other than the Gelber and McNamara study just discussed, is Laura Leets, Experiencing Hate Speech: Perceptions and Responses to Anti-Semitism and Antigay Speech, 58 J. Soc. Issues 341 (2002). See Brown, supra note 3, at 613 n.39. This study too suffers from the problem of not specifying if the hate speech was part of public discourse or personally directed speech. Of course, although not specified in the studies, it may be that some of the speech involved in these studies was public discourse that in turn deterred others from participating in public discourse. So under a very charitable view of what counts as empirical evidence, Gelber might be right that “[i]t is therefore not true that there is no evidence that silencing operates in the ways that defenders of hate speech laws allege.” Gelber, supra note 3, at 623 (emphasis added). Still, although defenders of hate speech laws have made this silencing argument for decades, the evidence that they have been able to produce in support of this theory is remarkably paltry.
. As I observed in my Opening Article, Brown had previously noted the objection that the silencing argument was supported by a “paucity of evidence.” Weinstein, supra note 1, at 579 (citing Brown, supra note 170, at 198 (2015)). Tellingly, however, Brown in this work did not express disagreement with this objection or cite to any evidence supporting the silencing effect. Instead, he sought to avoid this objection by invoking the precautionary principle. See Brown, supra note 170, at 199. Similarly, as Brown candidly acknowledges, see Brown, supra note 3, at 613 n.39, it was only after the editor of this Symposium suggested that he try to support the silencing argument with some evidence that Brown revised his Commentary by citing two articles which he asserted without analysis supported the silencing effect. Id. As I discuss in note 173, supra, if these studies supply any relevant evidence at all, it is at best meager evidence supporting Brown’s claim that hate speech prevents anyone from participating in public discourse. But equally as important, that Brown was on at least two occasions content to invoke the precautionary principle in the face of the objection that there was “a paucity of evidence” supporting the silencing principle without citing to any evidence of such a silencing effect, shows just how truly empirically insensitive his argument is. In note 229, below, I discuss Shiffrin’s interesting argument that the harms he claims hate speech causes need not be supported by empirical studies.
. I assume that those engaging in this “process of interpersonal justification and consensus” are concerned not exclusively with their own interests but also to some extent at least with the fair treatment of other groups in society as well as with the good of society as a whole. One of the many problems with Brown’s use of this “process” is it does not give us enough information about the parameters of this decision making process.
. See Brown, supra note 3, at 601 (emphasis added). Occasionally Brown acknowledges that such rejection will depend on circumstances; but more often this qualification is lacking, suggesting that it can be determined from the arm chair.
. The same objection applies to all other iterations of Brown’s use of “the process of interpersonal justification and consensus” to try to prove the propriety of the types of hate speech laws he supports. Besides invoking it in support of the silencing argument, Brown uses this process in his Commentary to make a more direct attempt to prove these types of hate speech laws are required by fundamentals of justice and thus promote political legitimacy. Brown thus argues that “free and equal people might reasonably look upon the adequate protection of their equal civic dignity, such as via group defamation laws (sensu stricto) or incitement to hatred laws, as a precondition of any notional agreement to joining the political community.” Brown, supra note 3, at 609. He adds that “[p]erhaps there are other fundamentals of justice, such as safeguarding people’s sense of their physical security, that is, freedom from legitimate fear of acts of discrimination or violence, that are also preconditions for any notional agreement to joining the political community, and that would also require laws, including incitement to hatred laws, that combat hate speech that contributes to a climate of fear.” Id. at 609–10. I agree that free speech doctrine so “aggressive” or “absolutist,” Brown, supra note 3, at 601, that it forbids speech restrictions needed to adequately protect civic dignity or prevent legitimate fear of acts of discrimination or violence will diminish the legitimacy of the legal system with respect to the subjects of hate speech. But as with the silencing argument, whether it would be reasonable to reject the failure of a legal system to enact the type of hate speech laws he supports as contrary to “fundamentals of justice” depends on a number of crucial empirical inquiries. The same goes for Brown’s argument, considered in my Opening Article, that it would be reasonable for members of minority or vulnerable groups to reject as contrary to fundamentals of justice the justification for not enacting hate speech laws that such laws may impair the legitimacy of downstream laws from which these groups benefit. See Weinstein, supra note 1, at 576–78. As I have emphasized, pertinent to this inquiry is the application of hate speech laws in actual operation to temperate criticism of homosexuality and Islam, a phenomenon that Brown largely ignores. See supra, text accompanying notes 135 to 142.
. In my Opening Article, I pointed out that there was an incommensurability problem in weighing the loss of systemic legitimacy that Brown claimed resulted from hate speech itself against the loss of legitimacy to particular laws that I claimed resulted from hate speech laws. See Weinstein, supra note 1, at 577. In his Commentary, Brown insists that his account of political legitimacy does have the “wherewithal” to respond to my claim about loss of legitimacy to downstream laws caused by hate speech bans. Brown, supra note 3, at 603. Brown writes that he would tell those whose speech was curtailed by hate speech laws that because the conduct prevented by downstream antidiscrimination laws is “clearly unjust” they have an obligation to obey these measures even if “to some extent the[ir] collective authorization” and “democratic legitimacy” has been reduced by the upstream speech restriction. Id. at 603–04. He adds that these people should keep in mind that these laws “curb forms of hate speech that can be corrosive of a shared, public sense of the basic elements of people’s equal status and dignity as members of society in good standing.” Id. at 604. Like all his uses of “the process of interpersonal justification and consensus,” whether those whose speech has been curtailed can reasonably reject this proposition depends on numerous empirical inquires, including once again the extent of speech curtailment and thus the amount of reduction of “collective authorization” and “democratic legitimacy” resulting from these allegedly “narrowly framed” laws in actual operation. Id. at 603–04. Aside from failing to address these crucial empirical questions, Brown’s attempt to justify these laws to dissenters has the added defect of talking past my basic critique of hate speech bans and political legitimacy. As I make clear, I share Brown’s view that the moral interest in preventing discrimination on the basis of race, religion, sex, and sexual orientation in “jobs, housing, transport, services, and so forth,” id. at 603, is so strong that it will be a rare case in which a speech restriction can put in doubt the morality of enforcing such laws. See Weinstein, supra note 1. The Evangelical Photographer scenario was meant to represent such a rare yet realistic case. Rather than engage this challenge, Brown simply ignores it.
. See id. at 604–05. Thus Brown does not argue that the failure to enact hate speech laws entitles “minority and vulnerable groups” “to rise up in revolution” (see Waldron, Political Legitimacy, supra note 4, at 332) because their civic dignity and sense of security is not adequately protected. Rather, consistent with my view (adapting Dahl) about viewpoint-based speech restrictions diminishing the “reservoir” of a legal system’s legitimacy, see supra note 43, Brown argues that the failure of hate speech laws makes the legal system “less politically legitimate than it could be.” Brown, supra note 3, at 610.
. In this regard, it should be noted that the scope of hate speech restrictions that Brown argues for is considerably broader than the one that Waldron defends. Thus in addition to incitement to hatred laws and group defamation laws (sensu stricto) Brown supports “regulations limiting the use of negative stereotyping or stigmatization, and perhaps even Holocaust denial legislation.” Brown, supra note 170, at 214. See also id. at 146, criticizing Waldron for “overlooking other kinds of laws that might also be said to protect the high and equal sociological status of vulnerable groups.” Brown, however, resists any tradeoff between the promotion of political legitimacy he argues results from hate speech bans and the damage to such legitimacy that I argue that such bans might cause. Rather, he asserts “that political legitimacy takes lexical priority over and, therefore, cannot be traded off against, the collective authorisation or democratic legitimacy of downstream laws.” Brown, supra note 3, at 600. I am not sure whether Brown means to confine this claim of priority to bar such tradeoff to cases involving damage to legitimacy of downstream laws or to extend this priority to bar any proposed tradeoff between the loss of legitimacy he identifies through the “process of interpersonal justification and consensus,” on the one hand, and the loss of systemic legitimacy that I and others argue can result from speech restrictions, on the other. But there is no need to pursue this question, for Brown merely asserts but does not argue for this priority. Relabeling the detriment to political legitimacy arising from speech restrictions that I stress in this Symposium as “collective authorisation or democratic legitimacy” while reserving the term “political legitimacy” for his account of legitimacy is surely not an argument for such priority. And I can think of no reason why the legitimacy concerns that Brown derives from a theory based in hypothetical consent should categorically take priority over legitimacy concerns that I and others derive from a theory of political participation based on formal equality.
. As my colleague Gary Marchant has observed, “there is no standard text” for the precautionary principle. Gary E. Marchant, From General Policy to Legal Rule: Aspirations and Limitations of the Precautionary Principle, 111 Ethics & Envtl. Health 1799, 1800 (2002). However, he notes that each formulation of the principle shares the prescription that “scientific certainty is not required before taking preventive measures.” Id. In contrast, he observes that there is great variance among various formulations of the principle in the level of threat necessary to trigger the principle. Id. Brown does not expressly adopt a particular formulation of the precautionary principle. But in discussing the evidentiary uncertainty of the effectiveness of hate speech bans, Brown writes: “According to what I shall call the Precautionary Principle, where the effects of doing nothing to reduce hate speech are sufficiently grave or serious, evidential uncertainty about what measures are minimally effective in reducing hate speech/the evils of hate speech ought not be used as a basis for not pursing measures that could be effective.” Brown, supra note 170, at 247. Note the potentially significant difference between lack of “scientific certainty,” which Marchant reports is common to all versions of the precautionary principle and Brown’s idiosyncratic reference to mere “evidential uncertainty.” The former standard means that there does not have to be conclusive proof to justify remedial measures, while the latter formulation would support, as Brown argues for, remedies based on even the most meager evidence. In his Commentary, Brown writes that there must be “at least some minimally adequate evidence that the relevant activities have certain effects and that these effects are potentially harmful in order to” apply the precautionary principle. Brown, supra note 3, at 613. He does not, however, specify what standards are to be applied to determine the evidence in question is “minimally adequate.” Rather he merely asserts “this threshold has been met for hate speech and various types of silencing effect,” citing without discussion the two studies discussed in note 173, supra.
. Specifically, those opposing the bans would have to produce evidence “sufficiently rigorous, comprehensive and abundant to command a consensus among the relevant body of experts” that the hate speech if left unregulated would not have a silencing effect. Id. at 613. In contrast, to invoke the precautionary principle, and thus shift the burden of proof, those supporting hate speech bans would have to produce only “some minimally adequate evidence” of the silencing effect. Id.
. See Weinstein, supra note 1, at 579–80. As noted in that article, id. at 580 n.186, Jill Hasday also suggested a plausible answer to this query, namely, that the number of Bs being silenced by the hate speech is greater than the number of As who will be silenced by the hate speech ban.
. I am assuming here that the intended (let alone effective) scope of the ban is not just on highly vituperative expression as Waldron imagines. Rather, in accord with bans actually on the books, I am assuming that the scope of bans are along the lines of those endorsed by Brown, which prohibit group defamation (sensu stricto), incitement to hatred and perhaps even Holocaust denial. See supra note 186. These restrictions plainly do not leave speakers free, even theoretically, to express the basic “propositional content” of certain views.
. Curiously, Gelber claims that I fail “to acknowledge what would happen to [my] argument if [I] were to concede that hate speech itself is capable of undermining the equal opportunity in decisionmaking” that we both agree is “fundamental to political legitimacy.” Gelber, supra note 3, at 624. In fact, I specifically addressed this situation in the scenario in which hate speech prevents members of an indigenous population from speaking in public discourse in favor of an exemption from drug law prohibition of a substance they use in their religious ceremonies. Weinstein, supra note 1, at 580–81. I acknowledge that in such a situation banning hate speech to promote the political legitimacy of downstream legislation would be justified if 1) the gain in legitimacy at least marginally exceeds the detriment to legitimacy caused by the speech restriction; and 2) there were no non-speech restrictive means by which government could ameliorate the “silencing effect” of the hate speech. (Both of these caveats stem from the basic liberal precept that the government always bears the burden of justifying the use of coercion, with the weight of this burden varying depending upon the various interests at stake.) I took a similar position with respect to resolving conflicting claims concerning systemic legitimacy. See id. at 580 n.186.
. See Jennifer Hornsby, Speech Acts and Pornography, in The Problem of Pornography (Susan Dwyer ed., 1995). See also, e.g., Tyler Kingkade, Texas Tech Frat Loses Charter Following ‘No Means Yes, Yes Means Anal’ Display, HuffPost (Oct. 9, 2014) http://www.huffingtonpost.com/2014/10/08/texas-tech-frat-no-means-yes_n_5953302.html.
. For instance, Gelber posits that homophobic speech in public discourse causes same-sex couples to become “fearful of walking down the street holding hands, and fearful of violent attacks against them and their property.” Gelber, supra note 3, at 622. She then asks whether this reasonable fear and knowledge of the experiences of other same-sex attracted people might mean that this “same sex couple could be silenced in much the same way” as the Evangelical Photographer in the scenario in my Opening Article felt unable to express her views about a pending bill to extend public accommodation antidiscrimination to include sexual orientation. Id. See also Stone, supra note 3, at 691–92. It is possible that even in some mature, stable democracies homophobic speech in public discourse reasonably deters same-sex couples from publicly supporting a law forbidding discrimination on the basis of sexual orientation in much that same way as I posited that hate speech laws deterred the photographer in my scenario from opposing such a law. But given the paucity of evidence supporting the silencing effect of hate speech, see supra note 173, including of homophobic speech, as part of public discourse on those wishing to engage in public discourse, it is difficult to make any meaningful assessment of this possibility. And as discussed above, Gelber’s own research does not differentiate between hate speech constituting public discourse and that which does not; nor does it distinguish between those deterred from participating in public discourse from other types of “silencing.” Id. Accordingly, it does little to support her “belie[f],” Gelber, supra note 3, at 622, that such silencing exists.
. Accord Robert Post, Cultural Heterogeneity, and Law: Pornography, Blasphemy, and the First Amendment, 76 Cal. L. Rev. 297, 326-28 (1988). For a detailed response to the silencing argument, including this variation, as it pertains to hate speech, see Post, supra note 92, at 306-17.
. Gelber, supra note 3, at 620–21. See also id. at 625 n.21. Gelber claims that like Ronald Dworkin in his debate with Langton, I “speak past” and appear “not to hear or recognize” these claims of the constitutive harm in hate speech. Id. at 622. But I am afraid that once again it is Gelber who does not “appear . . . to recognize” that my Opening Article was not meant to be a comprehensive discussion of the multifarious arguments for and against hate speech regulation (and it certainly did not intend to discuss the regulation of pornography). Rather, having dedicated the bulk of that article to considering how hate speech bans might deprive particular downstream laws of legitimacy, I felt it appropriate to consider also how hate speech itself might deprive particular laws of legitimacy. Not addressing every argument in the literature that might have some bearing on countervailing legitimacy concerns is hardly “talking past” or “not hearing” these arguments.
. Frederick Schauer suggests that the harm caused by hate speech might be constitutive of harm in much the same way as would the disclosure of a secret algorithm used by the Internal Revenue Service to determine whom to audit. Schauer, supra note 3, at 666. The analogy is inapt. The harm in the disclosure of the algorithm, and the government’s justification for prohibiting the disclosure, concern predominately, if not exclusively, the facilitative rather than the persuasive power of speech. In contrast, the harm in hate speech as part of public discourse, and the government’s reasons for prohibiting such expression, is usually concerned, at least in significant part, with the persuasive power of hate speech, such as its ability to persuade people to have false beliefs about members of minority groups or to “stir up” hatred against them which in turn might lead to acts of discrimination or even violence. One of the merits of Waldron’s argument in support of the narrow hate speech bans he supports is that the harm he emphasizes—causing members of vulnerable minority groups to feel insecure about their status in society—is not concerned with the persuasive power of speech in this sense. Nonetheless, even this harm is not caused by facilitative speech analogous to the speech in Schauer’s example. An example of facilitative hate speech analogous to that in Schauer’s hypothetical would be a list on a neo-Nazi website of the home address and telephone number of prominent Jews, along with the names and addresses of the school their children attend. For a helpful discussion of the distinction between persuasive and facilitative speech and the implication of this distinction for freedom of expression, see Thomas Scanlon, A Theory of Free Expression, 1 Phil. & Pub. Aff. 204, 211-12 (1972).
. Similarly, a ban on only the most vituperative forms of hate speech that leaves speakers free to express the basic propositional content of their bigoted views might in theory provide such a cabinable principle if it could be shown that vituperative racist expression is more harmful than vituperation in other contexts.
. Indeed, Brown’s idiosyncratic version of the precautionary principle bears an uncanny and troubling resemblance to Chief Justice Vinson’s perversion of the clear and present danger test in his plurality opinion in Dennis v. United States, 341 U.S. 494 (1951), a decision which upheld convictions of the leaders of the American Communist Party. See supra text accompanying note 62.
. As such it would through a different rationale bring us back to the notorious lack of protection provided anti-war protestors in the United States and other democracies during World War I. See Schenck v. United States, 249 U.S. 47 (1919); James Weinstein & Ivan Hare, General Introduction: Free Speech and the Suppression of Extreme Speech Past and Present, in Extreme Speech and Democracy, supra note 9, at 2–3.
. Shiffrin, supra note 3, at 477 (alteration in original) (quoting Steven H. Shiffrin, Dissent, Injustice, and the Meanings of America 77 nn.168–69 (2000)). Shiffrin acknowledges that these claims of harm “do not depend on empirical studies, but the experiences of human beings. When speech lacks a strong connection to the values underlying the First Amendment, at least in my view, a demand for empirical studies before regulation is not defensible.” Id. at 677 n.14 While I disagree with Shiffrin that a demand for empirical evidence is “indefensible” just because speech lacks a “strong connection” to free speech values, I agree with him that under such circumstances a lesser quantum of empirical proof is usually justified. This is why, for instance, commercial speech is properly regulated in some respect on less definite showing of harm than is public discourse. Where we disagree is whether hate speech in public discourse lacks such a strong connection. Finally, I share Shiffrin’s view that hate speech causes many of these harms. What I am uncertain about is the extent and frequency of several of these harms, a question that only empirical studies might help answer.
. See Dennis v. United States, 341 U.S. 494, 545 (1951) (“The defendants have been convicted of conspiring to organize a party of persons who advocate the overthrow of the Government by force and violence [by using] language reasonably and ordinarily calculated to incite persons to such action, and with the intent to cause the overthrow as speedily as circumstances would permit. On any scale of values which we have hitherto recognized, speech of this sort ranks low.”) (emphasis added) (quotation marks removed). In response to the argument that the government could easily squelch any resurrection that such advocacy might provoke, the plurality noted the harm that even unsuccessful attempts “create both physically and politically to a notion . . . .” Id. at 509). I agree with Shiffrin that in the United States the Ku Klux Klan far more effectively promoted “governmental illegitimacy” than did the American Communist Party. See Shiffrin, supra note 3, at 678 n.16. But in other mature, stable democracies the threat of communist subversion and thus promotion of “governmental illegitimacy” might be greater than it was in the United States. In any event, that the harms created by racist organizations might be marginally worse than those from radical political ones seems too fine a distinction to provide the protection that should be afforded all forms of dissent in a free and democratic society.
. For further discussion of the distinction between a slippery slope argument and an argument about the breadth of a rationale, see James Weinstein, A Constitutional Roadmap to the Regulation of Campus Hate Speech, 38 Wayne L. Rev. 163, 184 (1991).
. This does not mean, of course, that hate speech bans are not vulnerable to valid slippery slope concerns. For a discussion of the mechanism of the slippery slope problem of banning hate speech in the United States, see Post, supra note 92, at 315–17.
. My complaint about the breadth of rationales for suppressing hate speech is in one way essentially connected to a slippery slope concern. Embracing a rationale that will allow the suppression of speech that should be permitted in a free and democratic society will make suppression of such speech more likely. And indeed, in my view and those of others, European democracies, including Britain, have for decades inadequately protected free speech in ways beyond hate speech bans, a problem which seems to have worsened in the last few years. For an excellent discussion of this phenomenon, see Jacob Mchangama, Europe’s Freedom of Speech Fail, Foreign Pol’y (July 7, 2016),
http://foreignpolicy.com/2016/07/07/europes-freedom-of-speech-fail/. I am not sure what role hate speech bans have played in enfeebling the protection of speech in these democracies. I suspect the causation is bidirectional, with initial weakness in overall free speech protection allowing the enactment of hate speech bans which in turn further weakened the protection of free speech.
. As I have shown, however, even if Waldron is correct that the typical hate speech law is meant to target only highly vituperative expression such laws have been misapplied to cover temperate expression some of which is arguably not even bigoted. See Weinstein, supra note 1, at 552–61. Such misapplication will definitely interfere with the speaker’s and would-be speakers’ ability to influence “the shape of public opinion.”
. Id. at 656. For this reason I concluded that if such restrictions on vituperation could in practice be limited in the way Waldron supposes they are, the damage to normative legitimacy would be minimal, though perhaps not as trivial as Waldron suggests. See Weinstein, supra note 1, at 551.
. Shiffrin writes that I claim that racists have a First Amendment interest in expressing bigoted ideas as a way of “feel[ing] better.” Shiffrin, supra note 3 at 678 n.16 (citing Weinstein, supra note 1, at 551). Shiffrin is wrong that I was claiming here that racists or anyone else have a constitutionally significant interest in participating in public discourse so as to “feel better.” Rather, I was arguing that contrary to Waldron’s claim that the purpose of racist speech is to make vulnerable minorities unsure of their status in society, another reason that bigots might engage in racist rants in public discourse is not so much “to make minorities feel bad . . . but to make [themselves] feel better.” Weinstein, supra note 1, at 551. Interestingly, however, the emotive function of public discourse emphasized by the Court in Cohen would seem to recognize something at least akin to a First Amendment interest in participating in public discourse to “feel better,” including by using offensive or abusive language to express an idea.
. In a previous Symposium in which Blasi, Post and I participated, Post wrote of Blasi’s commentary on his target article: “Professor Vincent Blasi most generously catches the fundamental aspiration of my own work . . . .” Robert Post, Participatory Democracy as A Theory of Free Speech: A Reply, 97 Va. L. Rev. 617, 627 (2011). I feel the same way about Blasi’s generous effort to understand the “fundamental aspiration” of my Opening Article in this Symposium.
. Id. What such direct reliance on equal citizenship sacrifices, however, is the legitimating function of free speech, especially as it relates to justifying the government’s use of force to make people comply with laws with which they reasonably disagree.
. I have previously expressed concern about positing freedom of thought as the basis of American free speech doctrine. See James Weinstein, Seana Shiffrin’s Thinker-Based Theory of Free Speech: Elegant and Insightful, But Will It Work in Practice, 27 Const. Comment. 385 (2011).
. At least this is true of normative legitimacy, my primary concern in this Symposium. With respect to descriptive legitimacy, it may be that a speech restriction might for some individuals destroy any sense of political obligation they have to the legal system.
. In this way, the relationship between free speech and political legitimacy is not unlike the relationship between junk food and health. While a steady diet of greasy hamburgers and fries will over time likely have a deleterious, sometimes even fatal, effect on one’s health, fortunately for those of us who on very rare occasions like to eat a Big Mac or the like, one such meal, or even two or three a year, will for otherwise healthy people not lead to such baneful results. Unfortunately, another characteristic that both poorly-justified speech restrictions and consumption of junk food have in common is that one indulgence tends to lead to another.
. The one exception was Jeremy Waldron’s critique which doubted that legal restrictions on the “chaotic and unformed” public debate can have the direct consequences for legitimacy that I claim they can, Waldron, supra note 2, at 709, and simultaneously pressed the view that voting for one’s representative was the primary way that legislation was legitimized. Id. at 708. I agree that under such a parsimonious view about the relative importance of free speech to political legitimacy, it would be difficult to make the case that speech restrictions can seriously damage or even destroy the legitimacy of certain applications of downstream laws. But not only is Waldron’s view normatively unappealing, it also does not accurately describe free speech as it operates in contemporary liberal democracies. See Post, supra note 3, at 654–55.