Should Public Universities Guarantee Free Speech?

This article first appeared on Dorf on Law.

On Monday, UC Berkeley Law School Dean Erwin Chemerinsky delivered two lectures at Cornell on "Free Speech on Campus."

I'm not going to try to recap everything Dean Chemerinsky said. Instead, I want to focus on what he described as a gray area: the scope of the obligation of government--including public universities like UC Berkeley--to pay the added safety costs associated with speakers with highly controversial views.

Dean Chemerinsky recited a fact he has previously highlighted ( e.g., here ): When conservative columnist/podcaster Ben Shapiro spoke at Berkeley, the university spent $600,000 on security; it was prepared to spend even more for the "free speech week" that Milo Yiannapoulos was ostensibly planning.

Thus, Dean Chemerinsky identified what he described as an open question in the law that is also a problem. On one hand, if a public university where highly controversial speech is planned shuts out a speaker on the ground that security costs are too high due to the risk of violence occasioned by people who oppose the speaker's message, that amounts to a heckler's veto; on the other hand, universities have limited budgets. Dean Chemerinsky asked whether Berkeley would be obligated to spend millions if, say, Yiannopoulos planned and executed a "free speech semester."

Note that this problem is hardly unique to universities. Speakers and groups with extremely unpopular views impose similar security costs on state and local governments when they hold marches and rallies.

Must the government spend enormous sums to provide security for neo-Nazis who decide they want to march in some town? The 1992 SCOTUS case of Forsyth County v. Nationalist Movement seems to imply that they must.

Dean Chemerinsky recognized that this is a genuine policy dilemma to which there is no ideal solution. His solution would be to permit the public university to take security cost into account in deciding whether to host a speaker, but to place the burden on the university, if sued, to show that the decision to deny a speaker access was based on an assessment of actual expected security cost--i.e., that the added cost of security was not a pretext for official hostility to the speaker's unpopular viewpoint.

That may well be a reasonable solution, but it is not valid under Forsyth County , at least as I read the case. When I pressed Dean Chemerinsky about Forsyth County during the Q&A, he said that the case only decided that vesting discretion in a government official to decide how much to charge march organizers for the costs associated with their march was unconstitutional; the case did not reach the question whether charging speakers based on actual security costs would be valid.

With respect, I disagree with this characterization of the case. Forsyth County was a facial challenge to a permitting scheme that granted discretion to a government official to assess a fee for a permit to hold a march or rally. The Nationalist Movement was charged $100. The fee was invalidated on the ground that the scheme granted the official unguided discretion.

Conservative commentator Milo Yiannopoulos is escorted to the University of California, Berkeley campus to speak to dozens of supporters on September 24, 2017. Although a student group cancelled plans for Free Speech Week, Yiannopoulos was able to speak on campus surrounded by a heavy police presence. JOSH EDELSON/AFP/Getty

But the key reason why that discretion was problematic was the risk that it would be exercised so as to charge speakers more based on the increased security costs due to the unpopularity of their message.

If a licensing scheme is unconstitutional because it poses the mere risk of charging speakers based on security associated with hostile audience reaction, then surely a fee structure that explicitly takes account of those security costs--like the one Dean Chemerinsky proposes--would be unconstitutional a fortiori .

Dean Chemerinsky offered another reason why, in his view, his suggested alternative would at least fall into a gray area under Forsyth County : If the university did not itself make any determination based on its own disapproval of the speaker's message, he said, but only based on the prediction of security costs by a neutral third party (such as a police department), then there would be no censorship.

That's also a mistaken reading of Forsyth County in my view. There, Justice Blackmun, speaking for the Court, flatly said the following: "The costs to which petitioner refers are those associated with the public's reaction to the speech. Listeners' reaction to speech is not a content neutral basis for regulation."

Whether it's the university itself or the police department assessing the security costs associated with an unpopular speaker, the assessment, in Dean Chemerinsky's proposal, takes account of the public's reaction in exactly the way that the Forsyth County Court said is impermissible.

Having said that, I hasten to add that I believe Forsyth County should be limited. Very high security cost should be a valid reason for denying a speaker the opportunity to speak at some particular time or place, even if the consequence of such a rule is to allow and even encourage a heckler's veto.

That's not a good solution, but it seems to me more sensible than the alternative, in which an unholy alliance of provocative speakers and violence-threatening counter-protesters chews up a university's or locality's budget, crowding out other vital programs, including other speakers.

To fit that sort of limitation into free speech doctrine, a court could hold that, at a certain security cost threshold, saving the money needed to provide adequate security is a compelling interest.

The remaining question would then be what mechanism for allocating scarce resources should be chosen. Under these circumstances, perhaps it would be permissible to follow Dean Chemerinsky's proposal--charge the speakers for the added projected security costs as determined by a disinterested (but not "neutral" in the Forsyth County sense) third party.

Or perhaps Dean Chemerinsky's approach is still more content-based than necessary. One might think that a better approach was suggested by my colleague Prof. Kevin Clermont during the Q&A: At the beginning of the semester, the university would set a budget for non-routine security costs associated with speakers; student groups or others proposing to bring in such speakers would then be selected by lot, with speaking slots thus allocated randomly until the budget is used up.

Admittedly, this system would still be somewhat content-based: A speaker with only routine security costs (a lawyer lecturing on the legislative history of ERISA, say) would be automatically entitled to speak because no one will protest (or even attend!) her talk, while a speaker with non-routine costs (a sociobiologist arguing that women are unfit to be scientists, say) would be relegated to the lottery.

But this degree of content discrimination would be justified, in my view, on speech maximization grounds. Subjecting the routine-cost speakers to the lottery would greatly diminish the number of speakers.

I don't think that Prof. Clermont's solution is perfect, but there is no perfect solution here, because we have a genuine conflict between two valid propositions: (A) Hecklers shouldn't get to veto speech; and (B) Universities and governments shouldn't have to spend all their resources on providing security for controversial speakers. The best we can do is trade these principles off against each other in some reasonably sensitive way.

Michael C. Dorf is the Robert S. Stevens professor of law at Cornell University. He blogs at