Editor’s note: This article was adapted from a keynote presentation delivered at the National Law-Related Education Conference, on October 20, 2018, in Chicago. The theme of the conference was “Free Speech Today.”
I was asked here today to discuss whether the First Amendment is obsolete. In some sense, the answer is obviously no, in that it still has its basic job of protecting unpopular speakers from government oppression. In the news recently was a Louisiana city whose mayor tried to prevent volunteer clubs from buying Nike products and then using city facilities after Nike aired its Colin Kaepernick ads. The federal government just tried to prosecute a woman for laughing at Jeff Sessions during his confirmation hearing for Attorney General. So, the First Amendment still definitely has its classic job, but the question of whether our First Amendment doctrine, the rules that emerge out of the case law, is fit for purpose is actually a much harder question. There I think I have to say no. I want to talk today about two big problems, which I think are linked, although I think they can be distinguished.
Problem 1: Today’s First Amendment Doctrine Misses the Mark
Current First Amendment doctrine protects too much speech it shouldn’t protect and isn’t sufficiently concerned with threats to speech that don’t look like classic book burning or the police showing up and dragging you away for your speech. I think this is fundamentally a crisis of epistemology, a breakdown in courts’ understanding of how we know what we know. This can be seen with the U.S. Supreme Court’s insistence that money in politics isn’t a problem. From the Court’s perspective, if citizens distrust the ability of private parties to spend unlimited amounts of money promoting their causes and candidates so much that people lose faith in government, that’s the construction that the citizens choose to put on it, and not a reality, and not a problem for regulation to solve.
To illustrate my argument, I’m going to draw examples from something that we encounter each day: commercial speech. Commercial speech doctrine is the redheaded stepchild of First Amendment doctrine more generally. First Amendment lawyers spend most of our time talking about political speech regulation, but commercial speech doctrine is what we use to regulate ads and things related to ads. It is proving harder and harder to regulate basic out-and-out commercial fraud because of the way that doctrines are changing around free speech.
In particular, we’re in an age of judicial distrust of other fact finders, combined with perhaps unwarranted confidence in the court’s own ability to find facts or know the facts. The resulting product is a mix of relativism and certainty that tilts against the government’s ability to protect citizens and consumers from bad actors. In 2017, for example, Chief Justice John Roberts called the field of sociology “gobbledygook,” which was a double statement of contempt given the rigorous empirical model that he actually happened to be criticizing came from political science, not from sociology. It is not just a rhetorical stance. It has effects on outcomes. In a variety of consumer protection cases in recent years, federal judges have simply dismissed scientific conclusions about the weight of evidence in favor of saying that if there’s any evidence at all in favor of an advertising claim, the advertiser is allowed to make that claim, no matter what the balance of the evidence says. From an empiricist’s standpoint, this is nonsense.
The key point in these cases is a rejection of ways of knowing that don’t come from a courtroom clash of witnesses to individual events. Legislation and regulation work by taking lessons from aggregates of data. It is true that legislatures too often treat anecdotes like data, but hostility to statistically-derived knowledge inherently disadvantages the practices of legislation and administrative regulation. Not coincidentally, it also disadvantages the practice of aggregating individual wrongs in court, known as the “class action.”
First Amendment Lochnerism
A lot of scholars have identified a problem of so-called “First Amendment Lochnerism,” which means striking down economic regulations in the name of the First Amendment, using freedom of speech to prevent regulation of what is a fundamentally economic activity, even though speech is the method of carrying out the economic activity. It is named after Lochner v. New York, the much-maligned 1905 U.S. Supreme Court case that declared limits to the number of hours that bakers could work unconstitutional under the Fourteenth Amendment. We now understand that case to be about power, the judicial power in that case to prevent the state legislature from enacting laws that redistributed economic power because of supposedly fact-indifferent neutral principles. The new form of First Amendment anti-regulation Lochnerism is often premised on a different conceptual hook than pure power: it’s knowledge, which then mediates how power can be exercised.
In 2017, the Ninth Circuit struck down San Francisco’s attempt to get soda sellers to disclose that sugary drinks contribute to obesity on the ground that the disclosure itself was misleading, in that not everyone who drinks sugary soda is obese, and obesity may occur without drinking sugary soda. The city could require such a disclosure, the court’s reasoning went, if it could prove that there was a special, unique causal connection between soda and obesity, but it didn’t, so it couldn’t. Similarly, the Second Circuit has held in recent years that it is unconstitutional for the Food and Drug Administration (FDA) to prevent prescription drug makers from making claims about their drugs that the FDA hasn’t approved unless the FDA proves that the claims are false, rather than simply proving that the claims don’t meet the FDA’s standards for what counts as scientific proof.
Another decision in this vein includes a 2011 U.S. Supreme Court decision in Sorrel v. IMS Health, which struck down a state attempt to reduce the influence of drug detailers (including pharmaceutical representatives who visit doctors) and the gifts that they often bring on the prescriptions that doctors write. The Court ruled that “speech in aid of pharmaceutical marketing” is protected by the First Amendment. There are more of these cases at the lower court level, including a 2012 ruling that mandating graphic color photos of the health hazards of smoking on cigarette packs violated the cigarette companies’ rights to free speech.
There is a tension in these courts’ knowledge claims, and it’s this: a classic line from early in the Supreme Court’s history, from Marbury v. Madison in 1803, is that “it is emphatically the province and duty of the judicial department to say what the law is.” It’s almost as accurate to say that it is emphatically the province of the judicial branch to say what the facts are. That is, even if the fact-finding enterprise is occasionally shared with other branches, resolving most cases requires some sort of statement about their facts.
If we are truly in a state of deep epistemological uncertainty that the First Amendment demands, according to some of the rhetoric, then nobody in the government should be finding facts—courts also being agents of government. Taking an example from the Second Circuit drug scenario, a drug maker can make claims about a drug not approved by the FDA for which it has support in at least one decent study, no matter how much evidence there is against it. If it makes those claims, a patient takes the drug and the patient then suffers harm, then it’s perfectly possible that under current tort doctrine the patient can make a successful products liability claim against the drug maker based on the same standards of knowledge that the FDA tried to, but was not allowed to invoke, to protect its restriction on the drug maker’s promotional speech. It’s just going to be done by the jury instead of by the FDA. Yet one might expect that the FDA will usually do a better job of evaluating the evidence than an individual jury. In that case, we would have two holdings in severe tension, if not in direct conflict, because of their assessment of the state of knowledge about the drug. (Of course, First Amendment Lochnerites may also want the patient’s tort claim to be barred by the First Amendment—but that just shows how deeply they want to cut into the classic functions of government.)
The current strain of First Amendment libertarianism often denies other branches of government, coequal branches of government, the power to determine facts. If experts can’t agree, then there is enough uncertainty to allow the claim to be made; if you believe it, and it turns out to be wrong, tough luck. This position has the virtue of consistency, but it’s a blueprint for a truly terrible world, a nineteenth century world where businesses are essentially never liable for the harm they do to their customers or their employees.
Problem 2: Failure to Grapple with the Bad Faith Actors
The second big problem is also a crisis of epistemology, but it is different. It is a failure to grapple with the bad faith actors who use speech to create so much distrust and rage that ordinary communication and good faith debate breaks down. Theorists like Zeynep Tufekci in her excellent book Twitter and Tear Gas have suggested that the classic model of what counts as speech suppression is no longer descriptive of the biggest current threats to free and fair democratic discourse. In particular, she talks about how governments and other actors crowd out important information with trivia, pollute the information environment, encourage people to distrust news that doesn’t fit preconceived notions, and rely on groups of people acting in bad faith to create chaos and disbelief. Even when government encourages these things, our free speech doctrine doesn’t constrain government actors in the relevant ways.
Taking a recent example from the current administration, the courts have ordered President Trump to stop blocking Twitter users from his feed because of their viewpoints, but they haven’t ordered him to stop giving a megaphone to proven lies. It would be very hard to do so given current First Amendment doctrine and ideas about the relationship between judicial and executive power. In addition, I think that simply ordering him not to do this would be a really bad way of handling the problem. This is what makes bad faith in speech so wicked as a problem: because the solutions that we currently have on hand are just bad fits.
Speech Challenges and Powerful Online Sources
Relatedly, many current speech challenges are being resolved not by governments, but by ostensibly private platforms. They are still being resolved. It’s just a different set of actors ruling on what speech is okay and what speech will get a user banned. Yale Law Professor Jack Balkin argues that many of the dominant online players, Facebook and Google among them, should be treated as what he calls “information fiduciaries,” or trustees, toward their end users. That would mean that they have to exercise duties of good faith and non-manipulation of outcomes, and probably have some equal treatment and due process protections for users as well. For example, you’d have the right to know why you’re seeing a particular post, you’d have the right to contest a determination that your content shouldn’t be shown, and so on. This kind of regulation would be designed to promote free speech while also trying to limit unprotected and harmful speech, such as fraud and revenge porn.
The idea of information fiduciaries, while meant to foster free and fair public discourse, doesn’t look anything like conventional First Amendment doctrine. If anything, the platforms themselves are likely to protest that any such requirements would violate their own First Amendment rights to filter content or not filter content in any way they choose. Nonetheless, as online services become more like total institutions—places where we live so much of our lives—we, as consumers and as citizens, may want them to behave more like governments or branches of governments.
And they are behaving like governments already, just not like democracies. Facebook’s policy team has a biweekly content meeting. Different teams across the company, engineering, legal, content reviewers, and external partners provide recommendations to the content moderation team for inclusion in Facebook’s policy guidebook. The guidebook has examples of how much can you talk about beating a woman, for example, before it crosses the line to advocacy of violence against a particular woman. Noel Coward has a line in his play Private Lives (1930) that has been somewhat in the news recently that “certain women should be struck regularly, like gongs.” That meets Facebook’s standards for what it keeps up. When you’re doing content moderation on such a large scale, you have to have a list, and it’s a very distressing list.
Quite tellingly, the team leader calls this Facebook meeting a “mini-legislative session.” Professor Kate Connick has examined a bunch of the big services and she has found that Facebook is not unique. Everyone’s policies have marked similarities to legal or government systems with a detailed list of rules, trained human decision-making to apply those rules, and a reliance on a system of external influence to update and amend the rules.
But Facebook isn’t the only big entity affecting the information environment. As we have seen here and in other countries, governments and movements quickly learn how to use and misuse platform mechanisms for their own benefit. When government-backed teams are overwhelming social media with trivialities in order to distract from some important political event—this is common in China, where they have 100,000 people posting on behalf of the government—what policies and algorithms can identify that kind of pattern, much less sort the wheat from the chaff?
Balkin’s idea of fiduciary treatment would require platforms to make some sorts of attempts to fix these problems to protect users against fraud, or fake news, to protect users against the kind of bad faith claims that led conspiracy theorists in 2016 to assert that a pedophile ring backed by Hillary Clinton was operating out of a Washington, D.C. pizzeria. It is worth noting that when a guy actually showed up with a gun at that pizzeria, many of the same people who had spread the false claim condemned him. This is a really powerful indication of bad faith. If you believed that children were actually being held there, certainly, you would want someone to go to the pizzeria. If you were just using this factual claim as a symbol of how bad Hillary Clinton is, and you understand that, in fact, there was no pedophilia ring, believing (and perpetuating) the bad faith claim is wrong. Once speakers are willing to make these kinds of assertions, though, the possibility of actual communication about truth claims about the world seems to have gone out the window.
A question that has puzzled me about the idea of information fiduciaries is this: if speech policies within the major platforms are truly governance structures, why would we as citizens of a democracy accept anything other than democracy in the regulation of these spaces? If we truly thought about Facebook as a government, we would have to grapple with the fact that in the United States, the First Amendment means that a government can’t regulate much of the speech that makes parts of the web a hateful cesspool. I think many of us are willing to accept more interventionist models of content moderation because institutions like Facebook and Google are more like schools than like full governments. They are institutions that have incredibly powerful effects that are near total for some people some of the time, but there is still a much broader world outside where people interact. Schools, likewise, need to maintain sufficient connection to the overall democratic polity so that policies are, in the end, influenced both by all relevant stakeholders, but at the same time they can regulate a significant amount of speech, and the teacher can give you a bad grade for writing a bad paper even though a police officer couldn’t give you a ticket for making the very same bad argument in public. The public school system, even if it’s going to predictably fail in small ways, can be functional enough and better than the alternatives. That might be the best possible model for the larger information environment we now live in.
I know these examples have been depressing. I do apologize for that, but the one benefit of talking about epistemology at an education and First Amendment-related event is this: knowing how to know and teaching how to know are core elements of our jobs. This is our mission, and it is a huge one. If there is anyone who has the tools to re-establish effective and persuasive ways of knowing, it’s going to be educators.
Rebecca Tushnet is a professor at Harvard Law School. She has written extensively about First Amendment, copyright, trademark, and other related topics.