Free Speech and its Limits
These days, everything gets filtered through the pandemic prism of social distancing measures. Our discussions of civil liberties and human rights—topics that obviously predate the pandemic—now cannot even be had apart from situating them against our present context.
In general, this is good—we should do more to actively contextualize abstract and historical concepts to our present social circumstances. But one underexplored area to discuss in light of the pandemic is free speech.
We’re familiar with the discussions asking us to strike a balance between privacy protections and the collective good—such as when we consider how much privacy we should give up, if any at all, to enact meaningful contact tracing protocols. But what about free speech? Does the moment we now find ourselves in suggest a redrawing of parameters?
It is conventional to start a discussion about freedom of speech by considering John Stuart Mill’s principle on prohibiting harm.
What sorts of harms are at stake? Physical harm? Financial hardship? Emotional distress?
In On Liberty, chapter one, Mill writes:
The only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others.
This is an especially, perhaps ridiculously, strong claim.
There are certain libertarian schools of thought that take this statement to furnish us with a comprehensive governing principle, but most of us don’t think it’s quite enough to get an entire political philosophy off the ground. For example, how can the state fund the creation of a law enforcement body? How can the state administer regulation of common and individual ownership—a problem Mill and many of the English philosophers of the era considered?
Most political philosophy (I would argue, all interesting political philosophy) has come to regard the harm principle as governing free expression, as opposed to serving up a comprehensive governing principle. Using this more narrow interpretation of the principle, what are the limits it places on expression? When does expression cause harm?
To adequately answer this question, we need to explore what counts as a harm and what the relevant causal relation is.
The most stringent protection of free speech would not protect a man falsely shouting fire in a theatre and causing a panic. … The question in every case is whether the words used are used in such circumstances and are of such a nature as to create a clear and present danger that they will bring about the substantive evils that Congress has a right to prevent.
This is both a philosophical and substantive iteration of Mill’s harm principle. Causing a panic endangers, poses a significant risk of harm to, those who might be trampled in the attempt to escape. (Notably, Schenck v. United States holds that Congress could make it illegal for someone to distribute fliers encouraging potential World War I draftees to obstruct the draft … which is a contentious example, and an issue on which Holmes’s jurisprudence would evolve significantly.) Inciting a genocide is a standard example. If someone engages in rhetoric that one might reasonably expect to lead some members of the community to organize and engage in mass murder of an entire ethnic or religious group, then the rhetoric can appropriately be viewed as causing harm. After all, it causes the coordination and commission of mass murder.
Consider the act precipitating the lynching of Emmett Till: a white woman accused Till of threatening her, an accusation one could reasonably expect to result in the brutal murder of a young black man, given the social context. As such, her speaking was causally responsible (even if not solely responsible) for Till’s brutal murder.
These are obvious instances of harm. They are also cases in which speech bears causal responsibility for the tragic outcomes. While causal responsibility can get tricky (as we will see), one standard test is whether some harm would have occurred “but for” the relevant action. Would Till have been lynched without the allegation? Not plausibly.
What about some of the less simple cases? The challenge comes when it isn’t clear whether some individual is causally responsible, or whether someone is harmed, either on account of substantive disagreement about the definitions or because it is not clear whether the event satisfies an agreed-upon definition. The technical details of what might spur such a disagreement are interesting, but tangential.
Suppose someone’s speech is neither necessary nor sufficient for bringing about a harm, but the speech contributed to the likelihood and severity of a harm. Is that a part of the relevant sense of causal responsibility? My own inclination is to at least sometimes think this is enough, but this is where we get into some blurring, because there isn’t a clear boundary at which increasing likelihood starts to count as “responsibility,” among other concerns.
That doesn’t mean there are no clear cases. There are a bunch of problems around giving a definition of a “heap,” but that doesn’t mean every instance of “being a heap” is in doubt. Some things are obviously heaps, whatever the standard turns out to be.
With that in mind, I want to pivot to some questions in the recent range of free speech cases.
In the early days of the COVID-19 pandemic in the U.S., a handful of online hucksters and conspiracy theorists suggested that ingesting colloidal silver might be a way to inoculate oneself against the virus. Incidentally, those hucksters were also selling colloidal silver at the time.
Following these claims, state attorneys general in Missouri and New York and the FDA got involved. First, there was no evidence for their claim about colloidal silver’s efficacy; second, colloidal silver has a range of potentially serious adverse effects if ingested; and third, believing that they had been inoculated might motivate some people to engage in behavior that endangers public health. These legal threats shut down the hucksters’s sales pitch. So, why were the attorneys general so confident this was not a protected form of speech?
Making unsubstantiated claims is not always legally actionable. Even in cases where the unsubstantiated claims might arguably (morally) constitute a fraud and result in a financial windfall for a snake oil salesman, we allow a broad range of such claims. Faith healers are allowed to claim they can heal people for money; dietary supplement companies are allowed to make a fairly broad range of claims about the efficacy of their products, without substantiation. But the colloidal silver incident was not like those ordinary cases, for a few reasons. The most important reason is the incredibly high stakes associated with claiming to inoculate people against a (then incipient) public health threat.
The challenge with the colloidal silver claims is that, following the jurisprudence of Holmes, they were being made in a context of imminent and significantly elevated risk. COVID-19 is the crowded theater of public health.
It isn’t merely that these hucksters were huckster-ing. It’s that certain claims about infectious disease might influence a small group of individuals to act in a way that has a disproportionately large effect on the public health and welfare of the community. One lesson with which we are all too presently familiar is the relationship between a small community of folks behaving badly dramatically elevating the risk for large groups.
Consider the reasoning underlying the “fire in a crowded theater” metaphor. According to Holmes, the issue is that the claim of fire creates a “clear and present danger” not because it might create panic, but because (in the context of a crowded theater) that panic creates a high risk of folks getting trampled. It is not merely the risk to any one person—such as the speaker or a responsive listener—but to the collective welfare of the community.
An artifact of the liberal position is that there is a range of potential ways of conceiving of harm. Physical harms, like being trampled or being infected with a serious respiratory condition, seem uncontroversial enough. But there are other varieties of effects that may not obviously count as harm.
Being infected with a disease is a fairly uncontroversial case of harm. As are the side effects of ingesting colloidal silver. Even without the firm punch to the face, the physical damage incurred by these is such that no one really disputes that they are harmful.
But consider some alternative cases.
On the one hand, we extend especially broad protections in the case of political speech. The United States does this to expand broad rights to dissent, especially following an extensive history of cases like Schenck (where “fire in a crowded theater” was used to justify the suppression of political critics). In lots of circumstances this broad protection may come into direct tension with potential harm.
On the other hand, inducing fear for one’s life by presenting a plausible and imminent threat (even if only as an act of political speech) is a strong candidate for harm.
Is the potential threat directed at someone? In the case of the Michigan protesters, it seems fairly clearly directed at legislators and the governor, especially political actors who do not support reopening.
Is it meant to coerce action? The point of political protests generally is to motivate action; there’s a blurry line between coercing and politically motivating. There are clear cases. Burning a cross on a black family’s lawn is a clear instance of coercion; walking, unarmed, with signs is a clear instance of politically motivating without coercion. My own assessment is that bringing a big and clearly visible gun into the Capitol gallery is, at a minimum, analogous to the standard case of conversational implicature where someone on the street produces a gun and asks for money. We know what’s going on there.
Couple those considerations with the use of Nazi imagery at the Michigan protest and the subsequent recognition that certain lawmakers in particular require armed protection, and this pushes hard on the boundaries of even expansive free expression protections offered in political contexts.
I allude above to the problem of making sense of causal responsibility. When does speaking cause anything?
There are some easy cases. Under normal circumstances, saying “I promise” is sufficient for making a promise. (It’s actually constitutive, not strictly causal, but that’s another kettle of fish. It turns out, legal theories of causation include constitution, so it doesn’t matter for present purposes.)
In many cases, the impacts of any particular instance of speech are hard to measure. While statements from a leader (given the influence of leaders as … leaders) are more likely to directly and substantially influence actions, the statements of non-leaders may influence actions as well, but are more likely to do so indirectly, or by way of raising likelihoods, or by influence in aggregate with lots of other statements.
Consider a boring example. Sally posts on Twitter that the current unemployment rate is 14.7 percent. You don’t consider Sally an authority on the economy, so you’re reasonably skeptical of this claim. Sally saying “the unemployment rate is 14.7 percent” isn’t going to be sufficient for you to believe that the unemployment rate is 14.7 percent, or meaningfully change your behavior in any way. However, if Sally is making the claim in concert with a bunch of other folks (of varying levels of authority) making relevantly similar claims, that can definitely influence your beliefs about the unemployment rate. Sometimes this might not be enough to cause a change— but sometimes it is enough.
We collect lots of information now; we can see the things that people are saying much more readily, and so these judgments about the plausibility (or even truth) of a claim in virtue of the amount of repetition is significant.
This brings us back to a set of problems in philosophy about causal responsibility and leadership, but also the causal responsibility of non-leaders.
The late philosopher Derek Parfit gives an example of partial influence in a thought experiment. Suppose that there is a large community of individuals and each individual can press a button; in doing so, they contribute a small, on-its-own-unnoticeable amount of harm to a person. The community itself causes an enormous amount of harm when all, most, or many of the members push their buttons, but each individual is not substantially responsible on their own.
Aunt Betty shares a bogus claim about the preventability of the virus. On her own, Betty is unlikely to influence people; but suppose all of the world’s Aunt Bettys are reproducing those claims in various contexts. We may look on and say “Well, that’s just the amplification of nutty conspiracies and shouldn’t be taken seriously.” Maybe so, but some will take it seriously and will act in accordance with that claim.
It’s one thing for the American Medical Association to publicly criticize some for spreading misinformation about the virus, but one cannot plausibly control the Facebook and Twitter posts (and chain emails, lest we forget) of the general public, nor should we. The challenge is in assessing causal responsibility and working out how to address and curtail the potential harms that come about, whether or not we think the Bettys are morally responsible or that social media platforms or even public health agencies should address their wingnuttery.
I started the post with a formulation of the liberal position on free speech and expression. Now consider a “conservative caveat,” which is when there are certain domains where borderline cases ought to give strong deference to the person engaged in speech; that is, we ought to take a small-c conservative tack in restricting speech in those domains.
Historically, suppression and prosecution of political speech (in America and elsewhere) goes very badly, even when ostensibly given in the best interests of the collective good. Schenck is an illustrative example. There, the ability of certain political groups to protest the war was undermined by a claim that their speech “harmed” the war effort and the country. The standard intuition now is that freedom of speech, expression, and even association should be treated in this conservative way when it comes to matters of political expression. Much as some of us might be deeply emotionally gratified by doing so, we do not get to break up white supremacist associations just because their existence constitutes a general harm. The potential or actual harms have to be specific and targeted, rather than nebulous. In the borderline cases, the conservative caveat requires giving the benefit of the doubt to the political speakers.
These are difficult questions; they’re hard to assess from our present position and made far worse and more complicated by a range of facts unique to our public health situation. In many cases, representing yourself as an authority involves some level of duplicity. Conveying medical claims can provide imminent health risks, not just to the person who believes the claims, but others in the community who might be infected. Physically proximate associations (while sometimes necessary for certain forms of political expression, like picketing) are themselves partly constitutive of behavior dangerous to the public health. And so on down a list of complicating features. Part of what we have to do at present is encounter these questions and cases in their complexity, rather than simplifying for political expedience.