The liberal conception of free speech and its limits

ABSTRACT

Unfortunately, many people today see the regulation of lies, disinformation, hate speech, and fake news as an infringement of free speech, at least when such speech is ‘political,’ despite the damage that such speech can do. But this very protective attitude toward speech rests on a mistaken understanding of the role of free speech in a liberal society. The right to free speech is based on the liberal value of freedom, and as such can be no broader than freedom itself. And freedom has always been subject to reasonable limits in a liberal society. Indeed, while the principles of toleration and neutrality are often cited as supporting a broad interpretation of the right to free speech, they also tell us that certain limits apply to that right. We need not tolerate speech that encourages intolerance, and while government should be neutral between reasonable conceptions of the good, it need not be neutral between reasonable and unreasonable conceptions. These ideas form the framework of liberal society, and as I shall show, also provide a guide for understanding what kind of speech is protected in a liberal society and what it is not.

KEYWORDS:

• Democracy
• truth
• autonomy
• self-realisation
• republican liberty
• equality
• social media

1. Introduction

We live in the Trumpian age. Hate speech, misinformation, disinformation, ‘fake news’, and outlandish lies threaten the very existence of democracy. All these forms of speech are now amplified through channels of traditional media that used to be known for their high levels of integrity but, in a growing number of cases, abide no limits now other than profitability, and sometimes not even that. Even more consequentially, false claims and destructive rumours can now be amplified through the facilities of social media to poison the minds of almost everyone in society. As a result, the nature and limits of free speech are now the focus of heated political debate in a way that has not been the case since free speech became a tool for expressing opposition to the War in Vietnam, when the focus was not so much on whether particular statements were true, but on the extent to which such opposition could be expressed in crude, explicit terms, or contain information that while true, was claimed to violate national security.

But the political demographics of ‘who advocates what’ with regard to free speech has changed dramatically since then. In the 1960s and 70s, almost everyone on the left thought that free speech was a necessary and perhaps even a sufficient tool for taking society into a more progressive modern age, and that the repression of such speech was a key tactic for those who wanted to retard any movement toward the light. Today, the political views of those on each side of the debate are much more muddled. Those advocating unfettered or only lightly-fettered speech can be found on both the left and the right.¹ And those advocating more robust restrictions on the kinds of abusive speech that concern us today can be found on both the left and the right as well.² I suppose this could be taken as a sign that we have now got the balance just right, but I do not think this is correct. Indeed, the purpose of this paper is to identify some misunderstandings that have led more than a few liberals to think that they must defend those on the far right who are now weaponising free speech and using it to work toward liberalism’s demise. For thinking that free speech – even in just the political arena – means unfettered or only lightly fettered speech, instead of facilitating freedom, is actually powering our fall into authoritarianism.³

Part 2 of this paper discusses the source of the right to free speech, for there is surprisingly more disagreement about this than one might think. Identifying the source, however, is essential if we are to intelligently discuss what the scope of the right to free speech might be. Part 3 shows that once we discern the source of the right, we can explain why certain kinds of speech are not within its protection, not because they bump up against limits imposed by other principles, but because they were never within the coverage of the right in the first place. Part 4 moves on to how other principles (primarily but not exclusively the principles of toleration and neutrality) impose limits on what might otherwise be within reach of the right, for freedom of speech can be no broader than freedom itself, and freedom has always been subject to reasonable limits drawn from these other principles in a liberal society. It also explains how these limits should be applied, why enforcement of these limits is not an infringement of liberalism but an implementation of it, why political speech is subject to the same limits as other kinds of speech, and why our current meagre attempts to control false and misleading speech are not enough to curb the kind of abuses we are facing today.

Before I get on with this, however, I want to make some preliminary clarifications. First, in this paper, I shall be dealing primarily with the interaction between the right to free speech and the spread of hate and disinformation. I believe my argument will help resolve many other free speech issues as well,⁴ but I shall not discuss these here. For now, whether my theory is comprehensive enough to deal with a wider scope of free speech issues must remain an open question.

Second, for purposes of this paper, I will assume that foreign-based actors who are neither resident in nor citizens of the relevant nation-state have the same right to free speech as those who are either or both. I actually think that foreign actors do not have such a right. But foreign actors clearly have no greater rights than those who are residents and/or citizens. So my assumption will simplify matters here. If foreign actors do have lesser rights, or no right to free speech at all, the domestic suppression of foreign-sourced hate and disinformation would be unproblematic. Nevertheless, there would still be plenty of domestic-sourced problematic speech about, and probably much domestic repetition of problematic foreign speech as well. So the importance of the arguments I am making here with regard to the limits of free speech under liberalism would remain.

Finally, I want to emphasise that this is not a paper about what the First Amendment to the US Constitution does or does not permit, or about how we might interpret Supreme Court jurisprudence to come to some sort of coherent view as to the right to free speech under US constitutional law. Because much of the existing literature on free speech is about this,⁵ however, much of this literature will have only an indirect bearing on the arguments I am presenting here. My project is about the broader and more fundamental question of how we should understand the right to free speech as a matter of liberal political theory. There is some concordance between this and arguments about the proper interpretation of the First Amendment, and at one time there was a great deal, but this concordance has been rapidly diminishing as the conservative super-majority on the current Supreme Court uses an ever-expanding conception of the right to free speech to unwind decades of liberal regulation.⁶ Those who find my argument in this paper persuasive may find it helpful for explaining why the Court is currently moving in the wrong (that is, an illiberal) direction on free speech issues. But this paper is not intended as a road map for preparation of a legal brief. I will reference First Amendment decisions only when these provide particularly helpful examples of the wider issues of liberal political morality I discuss.

2. Where does the right to free speech come from?

2.1. The right to free speech and the fundamental principles of liberalism

One of the most crucial questions in determining what the right to free speech (or freedom of expression, the slightly more general term for the same thing) does and does not protect is where, exactly, does the right come from within the liberal family of political theories? This is important because identifying the source of the right – whether it comes from a single source or multiple ones or is in fact a separate and independent free-standing right, not derived from anything but rather a fundamental principle of liberalism in itself – will tell us a great deal about how we go about determining both the scope of the right and its limits. Of course, the precursors of any right can usually be found in many places. I myself have argued that despite conventional wisdom to the contrary, many important rights that uncontroversially form part of the landscape of liberalism can be independently derived from either the principle of freedom or the principle of equality.⁷ But there are many fundamental principles in liberalism besides freedom and equality, and even rights primarily based on freedom or equality may be shaped by other fundamental principles of liberalism, in whole or in part. Sometimes there may be a clear route to a particular right starting at one point or another; sometimes there may be routes from more than one fundamental principle that are equally direct; sometimes a particular right may be primarily sourced to one fundamental principle but still receive nudges from several others. In practice, all our rights are derived in part from all our more general fundamental principles, each of which lends at least a few of its genes to the ultimately more defined, practical product we call rights. After all, despite the fact that the fundamental principles of liberalism are by definition fundamental, there are often echoes of the same ideas in more than one. Indeed, if this were not true, it would be hard to see how our fundamental principles could form a coherent set.

But even though every right has roots in every or at least many fundamental principles, this does not mean that every right is not primarily an expression of just one. As a practical matter, the only way to have a set of rights that are consistent with our fundamental principles and consistent with each other is to think of the scope of any right as being found in one fundamental principle, even if other principles support a right of a similar scope, with limits on that scope then drawn from other fundamental principles. I recognise that it is easy to confuse something that is not within the scope of the right as instead being within that scope but subject to limits generated by other principles. Perhaps this is why most theorists ignore this distinction. Nevertheless, the difference between these possibilities is important. If nothing else because the principles we are trying to implement may be significantly different in each case, and if we do not know which of these we are to look to for definitive guidance, we are likely to rely mainly on our intuitions, and errors and inconsistencies in our conclusions are likely to occur. And if I am correct about this, the first step in defining the potential scope of any right is to identify the primary source from which it comes.

2.2. What is liberalism?

To do this, of course, one has to know what the fundamental principles of liberalism are. I have set forth my view on this elsewhere, and unfortunately, this discussion is rather lengthy.⁸ But a full recital and defense of my view as to the content of liberalism is not necessary here. What I will say here is that while the term ‘liberalism’ is often used to refer to a collection of views advocated by the moderate left, this is not how I am using the term. I am using the term in its broader, more fundamental ‘Children-of-the-Enlightenment’ sense. Understood in this way, liberalism includes many different substantive and sometimes incompatible doctrines of political morality, each of which nevertheless shares certain general ideas about the moral framework to be applied to the design and operation of social and political life. One can accordingly be a liberal and on either the moderate left or the moderate right as those terms are commonly understood or anything in between. And note that I am not unique in offering such a broad, encompassing, understanding of liberalism – it is not only how many liberals understand themselves, it is also how many of those who are anti-liberal define what they are attacking.⁹ The reason why liberals can all think of themselves as members of the same political family yet take opposing positions on some of the important policy issues of the day, you see, is that the fundamental principles on which all liberals agree are general concepts, not detailed conceptions.¹⁰ As such, these principles do not generate specific recommendations for either action or belief without further refinement, and there can be disagreements between liberals on how these fundamental principles should be understood.

What this means is that that while I will be criticising some well-known theories of free speech in what follows, I am not claiming that these competing theories of free speech are illiberal. With the exception of those on the far right who are weaponising theories of free speech in an attempt to undermine the existing liberal order, and most of these approaches are espoused by politicians and public figures, not philosophers, all the theories I shall discuss are liberal in their underlying aspirations. The proponents of these theories are simply making mistakes in interpretating and applying the fundamental principles of liberalism, leading them to promote theories of free speech that are not as consistent with these fundamental principles as they could be and therefore do not serve the values of liberalism well. But this is the kind of criticism that liberals raise against one another all the time in the course of debating how the fundamental values of liberalism should be instantiated. The claim that there is only one way for a society to implement its fundamental values is a claim that is antithetical to liberalism and in fact characteristic of perfectionist theories on the far left and far right.¹¹

Liberalism, accordingly, is a little messy. Adding to the complications here, no society is fully and completely liberal, either in its embrace of the fundamental principles of liberalism or in its success in effectuating those principles. All societies exist on a sliding scale between liberalism and illiberalism, and in most societies, there is currently an open, ongoing struggle between liberal and illiberal ideas. Blatantly illiberal conceptions of free speech (for example, the view that maintaining the integrity of the social order requires that no one be allowed to criticise the government in any way) are running around even in societies that are currently mostly liberal. But what I am trying to do here is address what free speech should look like in a society that at least aspires to embrace and does indeed significantly instantiate the fundamental principles of liberalism. By doing this, I hope to identify the dangers that liberal societies face from losing sight of what liberalism says about the scope and limits of the right to free speech. For by misunderstanding what liberalism requires with regard to speech, some liberals may be unintentionally playing into the hands of those who want to replace liberalism, broadly construed, with something else.¹²

2.3. A statement of fundamental principles

Having said this, I take the fundamental principles of liberalism to include at least the following: a commitment to toleration, neutrality, equality, and freedom (or liberty, the word philosophers often use to refer to freedom); a commitment to the separation of religious and political authority; the belief that all members of a political community should have an opportunity to participate in political decision-making under conditions of full information; the view that the purpose of public discourse and debate is to persuade others of the rightness of one’s position by resorting to arguments that one’s opponents could not reasonably reject, that the rule of law applies even to the rich and powerful, that empirical evidence rather than emotion and pre-moral ideology should have primacy in forming our reasons for action and belief, that punishment should be informed by the principle of proportionality, and finally, that the individual, not the community, is the fundamental social unit, the point to which any search for moral responsibility must go and where the ultimate focus of our moral concern must lie. One could argue, I suppose, about whether each of these ideas is as fundamental as the others. If one begins with a general enough idea of one or two of the principles I have listed, it might be possible to derive all the others from these. Indeed, this is exactly what some theorists try to do using the principles of toleration and/or neutrality. But I think it is fair to say that any conception of liberalism that is worth taking seriously must identify each of the items on this longer list of ideas as playing a significant role in the shaping of society, regardless of whether they could be reduced to an even smaller set of fundamental principles by generalising the conception of some other fundamental principle even more.

I will make just a few more comments on the nature of the various principles I have set forth as forming the fundamental basis for liberalism before presenting my own understanding of the liberal conception of free speech. First, one critic of this paper has questioned why ‘democracy’ is not on my list of fundamental principles. This is a concerning omission, according to this critic, because a number of important theorists source the right to free speech (partially, if not fully) in some conception of democracy and how it can be most effectively instantiated and maintained.¹³ But democracy is on my list. I have just phrased that value in a more general way to make it clear that the kind of democracy that liberals embrace requires that citizens be provided with the opportunity to participate in political decision-making under conditions of full information. Clearly, not all self-proclaimed democracies do this – indeed, most states that liberals would consider authoritarian dictatorships are technically democracies.¹⁴ My phrasing of this value also leaves open the possibility that there may be forms of liberal social and political organisation that are not what we would call democracies, but which nevertheless do take the underlying value of public participation in self-governance under conditions of full information seriously.

Some readers may also wonder why I have not included ‘truth’ on my list of liberal values, as in the view that ‘open discussion and a free exchange of views’ is the best possible engine for the discovery of the truth in ‘the free marketplace of ideas’.¹⁵ Undeniably, this is one of the oldest and perhaps still most popular bases offered as the source of the liberal right of free speech.¹⁶ In fact, this idea goes back to J. S. Mill,¹⁷ although Mill himself did not use the ‘marketplace of ideas’ metaphor – this came later in the US Supreme Court’s decision in Abrams v. United States.¹⁸ But Mill said things that are very similar, and most of Mill’s followers think Mill would endorse the metaphor even though he did not use it himself.¹⁹ While the word ‘truth’ does not appear on my list of fundamental values, however, I have not left the pursuit of truth off my list. On my list, this value is expressed as the view that ‘empirical evidence rather than emotion and pre-moral ideology should have primacy in forming our reasons for action and belief’. So I do not deny that providing an effective mechanism for the discovery of truth is an important aspect of liberal political morality; I simply focus on what makes for an effective route to the truth rather than assume that the freer the speech the more truth will be discovered. That claim, and the associated claim that this makes the pursuit of truth the primary source of the right to free speech – I will address in a moment.

Autonomy and self-realisation (usually but not always grouped together) are also values that some readers might believe are missing from my list of fundamental liberal principles, and these values are also often cited as the progenitors of the liberal conception of free speech in one form or another.²⁰ But again, I have not overlooked these potential sources of the right to free speech. These ideas are embedded not only in some possible conceptions of the principle of freedom but also in the principle that ‘the individual, not the community, is the fundamental social unit’. But as I shall explain later when I discuss the work of Thomas Scanlon and C. Edwin Baker, autonomy and self-realisation can be understood in a variety of ways, and some of these seriously transgress other principles on my list, such as toleration and neutrality. One must accordingly be very careful about what one means when discussing autonomy and self-realisation – glib generalisations about their importance as values can lead us seriously astray. Yet they are currently bandied about by those who cite them as the source of the right to free speech as if they had established meanings and no problematic implications. But again, more on this later.

The main point I want to make now is that my list is intended to be reasonably comprehensive and uncontroversial. I am trying to get as close as possible to first principles without becoming so general that the list no longer conveys any practical information, or so specific that it suggests that liberalism is really just a form of moral particularism, a large, unsystemisable collection of customised individual principles that are each designed to apply to only one or very few practical cases, rather than a set of general principles from which particularised principles can be consistently derived in a way that makes them share important features and hang together as a coherent set. This means I will be operating at somewhat higher level of abstraction than some theorists do, while operating at a somewhat lower level of abstraction than others.

2.4. What is the source within liberalism of the right to free speech?

In any event, various liberal theorists have attempted to source the right to free speech to one or two and even more of these values and interests derived from them, even if they often use different terms to refer to some of them.²¹ The three main competing bases for the liberal right to free speech in the existing literature, in fact, are democracy, truth, and autonomy.²² Oddly, this leaves the most obvious source for the right of free speech – freedom – out. But other sources of the right are often offered as well, including some form of freedom, as well as many other actual and supposed liberal values. Indeed, many of the recipes presented for the liberal right of free speech contain complex, multi-factorial and somewhat romanticised expressions of their ingredients.²³ The literature on free speech is too vast to mention all of these, but here are some of them: free speech is primarily aimed at ensuring the self-realisation of the individual within the bounds of liberty;²⁴ it is designed not only to protect liberty, but also to affirmatively sponsor the individualism, the rebelliousness, the anti-authoritarianism, and the spirit of nonconformity within us all;²⁵ the right to free speech is best encapsulated by what we might call the ‘thinker-based approach’, which is itself based on a conception of personal autonomy and self-realisation²⁶; the right to free speech is best understood as containing a precise balance between the promotion of democratic self-governance and the prevention of harm to others;²⁷ it expresses a respect for autonomy, the dignity of human beings, personal security, privacy, citizenship and equality all rolled into one²⁸; it contains an even more multi-faceted combination of assuring individual self-fulfillment, advancing knowledge and discovering truth, providing for participation in decision-making by all members of society, achieving a more adaptable and hence a more stable community, and maintaining the precarious balance between healthy cleavage and necessary consensus.²⁹ The title of one recent work even claims that the right to free speech is an expression ‘self-restraint’, which sounds like it might focus not on the right to speak but on the duty to stay silent, but the argument presented is actually that the right to free speech is designed to prevent the derogation of ‘warranted self-respect’.³⁰ And so on.

Those that source the right to free speech to two or more of what I have identified as the fundamental values of liberalism I reject for the reasons I have already stated – the scope of a right must be sourced to one fundamental principle. Other principles may redundantly support a right of a similar scope, but this should not change our view as to which fundamental principle is primary. The most important function of other principles is to describe what limits apply to what would otherwise be the scope of the right to ensure that this right will not end up being a source of conflict as to rights generated by other fundamental values. And at least some of the theorists I have noted above do indeed claim to source the right to just one fundamental principle, with other principles being used to provide limits. At least these theorists can be read this way even if they do not make this approach explicit, which they typically do not do. In any event, I will get to these arguments in a moment. But for now, I want to point out that at the very least, regardless of whether one thinks the fundamental values of liberalism can be reduced to a single value or two or can be expanded beyond those I have listed, it should be clear that it would be odd to think of freedom of speech as a stand-alone right under liberalism – something meaningfully separate and independent from the fundamental principles that I have identified and of roughly equal status to them.

Some theorists, however, do make exactly this claim.³¹ But the argument for this is made using an analytically suspect method. That is, it begins by noting what are effectively the pre-conceived elements of the right to free speech, derived mostly through intuition. Then a (sometimes exhaustive sometimes superficial) search is conducted for where these elements might come from. The conclusion is then reached that freedom of expression must be a stand-alone independent principle, for at least some of its (very important) elements allegedly cannot be sourced anywhere else.³² But this method of reasoning begs the question at issue, which is what the scope and limits of the right to free speech might be within the liberal conception of political morality. The more defensible approach is to address this question by moving forwards instead of backwards. In other words, not by assuming the content of the right to free speech and then looking for its sources, but by beginning with principles of liberalism with which almost all liberals would agree, and then looking to see what kind of right to free speech these principles would produce. Indeed, proceeding in any other manner is most likely to lead us to an ad hoc collection of what are really just personal preferences, selected and mixed together from various fundamental principles and even non-fundamental ideas to produce a set of bespoke rights that are in reality chosen for pre-moral ideological reasons and not because they are derived from that pre-existing coherent set of fundamental principles we call liberalism. Moreover, this approach effectively mocks the idea of freedom by denying that it is the source of what is commonly understood as one of its most important constituent parts. Doing this would be more characteristic of an illiberal society, where a right is simply whatever those in power say it is, or exactly the kind of situation that liberalism is supposed to reject.

2.5. Freedom and democracy

But if freedom of speech is a derivative right, not a free-standing one, is it really primarily derived from freedom and not some other fundamental principle? I have already noted the importance of free speech that some theorists attach to the maintenance of a meaningful and effective form of liberal democracy. Some of these theorists even identify democracy as not just a partial but the primary source of the right to free speech.³³ And it is easy to see why freedom of speech is essential to the proper functioning of a liberal democracy. On the one hand, democracy is pointless unless dissenting and minority voices can be expressed and given the opportunity to convince others to change or moderate their current positions. If people were allowed to hear only views that they already endorsed, or only those that the government endorsed, then there would be no need for a political process, for everyone would already and always be on the same page. On the other hand, hate speech and disinformation erode the confidence we have that democracy will produce a carefully considered and informed choice of the people as to the composition of their government and the policies it will pursue. If we want to protect the democratic process, we want the public to be informed about the issues on the table, not misinformed, and we do not want their emotions to be so inflamed that they lose the ability to think and act rationally.³⁴ So the concerns based in democracy cut both ways. As a result, our conception of democracy might be a good source of limits on free speech – indeed, this is exactly what I shall argue in a moment. But it is not at all clear that the scope of the right to free speech would be any smaller if we thought of it as derived from our conception of freedom rather than some conception of what makes for a functioning democracy. Indeed, many kinds of speech would not have any clear relevance to the proper functioning of democracy whatsoever, but we would not want to exclude them from protected status merely because of that, for freedom of speech would not be much of a freedom in that case. So in terms of deciding the primary source of the scope of the right, the value of democracy does not seem like a good candidate for primary parentage, even if it might be an important source of limits on the right. This means that freedom remains the most obvious source if we are trying to see what the scope of the right potentially might be before we apply limits taken from other principles.

2.6. Freedom and truth

There are even more problems if we treat the pursuit of truth in the free marketplace of ideas as the primary source of the right, despite the undeniable gravitas of the theorists who suggest this.³⁵ Obviously, a lot of speech is false. So we cannot take the idea that free speech is there to protect true speech too literally or this would justify mass suppression. It is the backing and forthing of different claims and ideas that is supposed to result in the production of truth, for once brought into the light, the thought is that false claims will be revealed as false and only the truth will be left standing.³⁶ Accordingly, the best antidote for false speech, it is often said, is counterspeech, not regulation.³⁷ But context here is important. Remember that the free market metaphor for the production of truth came in when this was popular as a way of describing the benefits of a free market economy. Since then, we have learned that the supposed efficiency of an unfettered economy can be perverted in a great many ways by market failures. Regulation to correct these failures is therefore essential to keep the economy working efficiently and in the public interest.³⁸ Clearly, some market failures also exist in the free marketplace for ideas.³⁹ Why should we not want to correct for this if maximising the production of truth is our ultimate goal?⁴⁰

Moreover, when Mill was writing and for a long time thereafter, the concern was primarily the protection of dissenting voices, for in practice these were the ones most likely to be articulating truth in opposition to established falsehoods as liberalism took hold and matured.⁴¹ And it was difficult for dissenters to get their message out to the wider public, for the major platforms available for the dissemination of speech were largely controlled by those already in charge.⁴² The voices of dissenters were therefore easy to shut down, especially when the only readily available outlet for their speech was a literal soapbox. Now, however, circumstances have changed. Using social media, almost anyone can speak unvetted to a vast audience with ease, and this has led to the proliferation of conspiracy theories, lies, and disinformation. Even if this has made it easier for truthful information to be spread as well, it is hard to see the net result of this phenomenon as maximising the production of truth. For there are two ways to conceal the truth. One is to suppress it, the most common tactic used even after the appearance of liberalism and for a long time thereafter; the other is to bury it under so much other, irrelevant, false, and misleading information that its existence and importance are never discovered. Indeed, the latter may be the more effective and less risky strategy whenever it can be employed. It takes a great deal of discipline and resources to dig through the rubbish and discover the truth; most people lack both. So the goal of maximising truth may indeed still be relevant in a liberal society, but the usefulness of the ‘free marketplace of ideas’ metaphor for guiding us toward the truth clearly needs to be reconsidered.

Indeed, the connection between unfettered speech and the production of truth is an empirical claim, but empirical evidence is never offered to support it. What is offered in anecdotal evidence, usually about the suppression of liberal voices.⁴³ But these examples are not representative. If the weight of anecdotal evidence is to be the decisive factor here, the evidence of the dangers of unfettered illiberal voices is at least as great if not greater. Lies and propaganda, and even sincerely but unreasonably held beliefs about particular individuals and groups of people, as well as about the state of the world and human nature, have repeatedly been used to foment racial, ethnic, and religious division, undermine the bonds that hold society together, and eventually provoke widespread violence and even genocide. Just consider, for example, the rise of the Nazis, and the foment of racial hatred and internecine warfare in Rwanda and the former Yugoslavia.⁴⁴ The suppression of truth can lead to violence and destruction too, of course, but in states where some form of liberalism, even if not fully realised, has already acquired an established foothold, this is at least arguably more individualised and less widespread. On balance, it is the failure to address the proliferation of false and otherwise unreasonable claims in liberal democratic and emerging liberal democratic regimes that have repeatedly incited an often violent but in any event socially destructive political regression, and this is what puts liberalism at the greatest risk.⁴⁵

This is precisely the insight, in fact, that drives the activities of the Russian Internet Service Agency, the Russian government agency in charge of spreading disinformation and division in liberal democracies,⁴⁶ an insight that has also now attracted a growing number of domestic and well-financed foreign imitators (including the Chinese) and inspired a ‘billion-dollar disinformation campaign to reelect [President Trump]’.⁴⁷ In other words, maximising truth through rigorous debate only works if the framework that establishes the marketplace for ideas is itself equipped with protections to ensure that fraud, deceit, misrepresentation, and manipulation on an industrial scale is prohibited. Flooding the marketplace with so much noise and misinformation that it becomes impossible to pick out the truth, or what is most likely to constitute the truth when there is reasonable uncertainty, is just as much a violation of this liberal principle as suppressing the truth directly. Accordingly, if the production of truth were our primary concern, we have just as strong a reason to regulate the market for ideas as we have to leave it largely unfettered.⁴⁸

Take the common law adversary system as an instructive example. Like free speech in general, the trial has also been called ‘the greatest engine for the discovery of truth ever invented’.⁴⁹ The rules of evidence, however, strictly control what lawyers and witnesses can say and what documents and other forms of information can be offered in court and used by the finder of fact to make its decision. These rules still assume that the production of truth is maximised through the adversarial backing and forthing of each parties’ legal champions, but these rules do not simply assume that truth will arise out of a free for all. Instead, they adopt a very careful, regulated approach, and most people would say that this approach produces far more truth than if the presentation of evidence and argument were more unfettered. In light of this, the weight of anecdotal suggests that the best way to maximise the production of truth is not to simply to expect it to arise out of the marketplace of ideas, no matter how corrupt or chaotic, but to insist that we regulate speech in that marketplace in appropriate ways.⁵⁰

This becomes even more clear when we consider that on many matters, especially with regard to public policy and culture, there is not and probably never will be something equivalent to a ‘truth’. At least there are a great many important subjects for which there is nothing remotely close to an objective truth.⁵¹ Ensuring that truth prevails is thus clearly not a value that can do the work required in all cases on its own, and some of these are among the most important. Once again, it seems like this liberal value is more likely to make sense as a source of limits on the right of free speech, while the liberal conception of freedom would be the more apposite value for sourcing the right.

2.7. Freedom and equality

Even though we have disposed of the most popular arguments for the basis of the right to freedom of speech, there is yet one more to consider before we default to freedom. The highly-respected liberal philosopher Ronald Dworkin famously argued that the right to free speech is not actually derived from freedom, but from equality. Dworkin makes this move, however, because he takes freedom as referring exclusively to what philosophers call ‘negative liberty’, the extent to which we enjoy freedom from interference by other human agents. Dworkin argues that negative liberty cannot be a right, for if it were a right it would conflict with equality and almost every other right imaginable. This, in turn, would undermine our very concept of what it means to have a right, which is to enjoy an indefeasible zone of personal autonomy within which we can make certain decisions, regardless of whether these decisions are bad for ourselves or others or even the common good.⁵² And if negative liberty cannot be a right, Dworkin argues, it cannot be the basis of the right to free speech, for freedom of speech is a right, an individual right, something that imposes a duty of non-interference on identifiable individuals and not just a general floating social duty to remedy a systemic wrong as violations of distributive justice often (but not always) do.

Despite the convincing nature of Dworkin’s argument, however, some theorists do continue to assert that the basis of the right to freedom of speech is negative liberty. But they do so by simply ignoring Dworkin’s argument about the nature of rights, not by addressing and rebutting it.⁵³ Nevertheless, this does not mean that Dworkin’s sourcing of the right of free speech in equality is necessarily correct. For liberty has many different forms. Dworkin focused exclusively on negative liberty, but this is not the only possible conception of liberty that we could consider if we were looking to the source of the right to free speech in the liberal principle of freedom. Another possibility is ‘positive’ liberty, which is meant to refer to the degree to which one has managed to self-actualise, to realise one’s full potential.⁵⁴ Conceptions of positive liberty, however, often go beyond a conception of what is good for a particular person – conceptions of positive liberty are often political. That is, they include the view that it is government’s role to determine which conception of the good is best for a person and take steps to ensure that society is composed exclusively of persons who all embrace and generally abide by that particular conception. In other words, a political conception of positive liberty is a detailed vision of how everyone, not only oneself, should live their life, and includes the view that it is the role of government to ensure that everyone complies with this particular vision.

Of course, there are a great many such conceptions. It may even be the case that no two people share a conception of positive liberty that are exactly alike. Many of these conceptions will contain some version of the right to free speech, no doubt, but the versions embraced by each conception would vary greatly. While some people might embrace a conception of positive liberty that comes with a right to unfettered speech, some might embrace a conception that does not allow criticism of the government or the expression of any idea of which the government does not approve. Everything in between these extremes could also be found in one conception of positive liberty or another. And many other aspects of each conception would also vary widely. Some of these might even seem attractive to liberals.⁵⁵ But some, I am sure I need not remind anyone, are very scary indeed.⁵⁶ Choosing a single conception of positive liberty on which society as a whole should rely is therefore going to be very controversial.

To resolve this potential source of conflict, liberalism embraces the idea that society is characterised by reasonable pluralism as to appropriate conceptions of the good. As well as the principle of liberty, liberals therefore believe in the fundamental principle of neutrality, which says that rights should not be based on controversial comprehensive conceptions of the good, as well as the principle of toleration, which says that the government should not favour one reasonable conception of the proper way of life over other reasonable conceptions.⁵⁷ And this means that a comprehensive conception of positive liberty, like negative liberty, cannot be the basis of a right under liberalism, including a right to free speech, no matter what its form. For if it were, the government could oppress anyone who does not follow the government approved view, effectively using its own conception of positive liberty to compel others to be free in exactly the same way, no matter what their personal beliefs or desires might be. So Dworkin was correct in not considering a conception of positive liberty as a source for the right of freedom to speak within the realm of liberal political morality.

2.8. Freedom as republican liberty

But there is yet another conception of liberty that is a plausible candidate for sourcing the right to free speech that Dworkin did not consider. This is the conception that has come to be known as ‘republican’ liberty. This kind of liberty refers to the extent to which one is free from domination by the arbitrary will of another.⁵⁸ Dworkin did not consider republican liberty as a potential source of the right to freedom of expression simply because it had not yet been recognised as a separate and independent conception of modern liberty when Dworkin was writing about rights in the late 1970s – this did not happen until the very end of the last century. And republican liberty was not really treated a something worth separate consideration by a wide range of political philosophers until the early part of this century. Moreover, at the time Dworkin was writing, the conflict between the left and the right in general was viewed (and often still is) as a conflict between liberty and equality. Those who took a left-leaning view of rights sourced them in equality; those who took a right-leaning view sourced them in liberty, and this was true of the right to free speech as well.⁵⁹

For some time now, however, I have been engaged in an effort to reclaim the argument from liberty for the left.⁶⁰ This paper is another step in that project, and I will get to some of the elements of my argument on this in a moment. But the bottom line is that in my view, when the principle of liberty is properly understood, liberty and equality do not conflict; they lead to the same outcome. There is still much resistance to my view given the amount both sides have invested in the intellectual infrastructure of the liberty v. equality framework, but one advantage of re-framing the debate as not between equality and liberty but between liberty properly understood and liberty misstated is that liberty no longer remains a trump card for producing illiberal outcomes, as so many on the political right contend in a wide variety of areas.⁶¹ And given these developments, instead of engaging in the intellectual contortions that Dworkin had to make in order to get the right to freedom of expression out of equality, we can simply derive the right to free expression from liberty in the form of republican liberty directly. This does not mean that free speech is not also an important tool for the instantiation and maintenance of equality. The protection of equality may indeed provide yet another justification for imposing various limits on speech.⁶² But we do not have to run away from the idea of liberty in order to explain what free speech is and when and why we should protect it. For these purposes, there is no need to look to equality at all.

2.9. More about republican liberty

Can we say more precisely, however, what we mean by republican liberty? The best example of a complete infringement of republican liberty would be slavery, for the slave exists totally subject to the arbitrary will of his master. But lesser violations of republican liberty are possible too and, at least today, are much more common than the condition of slavery. Many conceptions of republican liberty that are currently popular today, however, are very thick – that is, in the course of expounding what might constitute domination, these theories go well beyond prescribing things that could be characterised as subjecting someone to arbitrary treatment and end up being almost entire theories of justice in themselves.⁶³ As inspiring as these theories may be to those who agree with the particular content included in them, the thicker these theories are, the more they look like comprehensive conceptions of positive liberty rather than something separate and distinct. And the thicker they are, the more controversial they become, and therefore the more likely they will stray into territory that violates the principles of toleration and neutrality.

The conception of republican liberty I will rely on here, however, is very thin. It does not rely on some interpretation of what ‘domination’ might mean apart from arbitrary treatment, as many other theories of republican liberty do.⁶⁴ And it does not expand the concept of arbitrariness so much as to make any kind of unfairness prohibited. For example, if one expands the concept of arbitrariness to cover ‘structural injustice’, ‘discursive injustice’, ‘social injustice’, or all three, as some theorists suggest we do,⁶⁵ this might even suggest that men should never be allowed to speak because, as some feminists point out, men cannot do so without their speech being given more weight than the equivalent speech by women in light of the biases and prejudices currently embedded in our social practices.⁶⁶ Realising this, the proponents of such a comprehensive view of republican liberty view temper their remedial suggestions – instead of prohibiting speech by men, they suggest that we enact provisions that simply encourage people to listen to women.⁶⁷ But I do not see why such moderation would be required by this conception of republican liberty – indeed, it seems like the argument for more drastic remedies would be strong, and this leads us toward a world in which the right to free speech could be used mostly to shut people up, which seems rather Orwellian.⁶⁸ In any case, the problem is that what is being used here is a very thick conception of republican liberty, one that looks much more like a full-fledged comprehensive conception of positive liberty, a view as to how to ensure that everyone can be the best that they can be. And this would violate the principle of neutrality and toleration, even if it might make society ‘better’ in the view of those who embrace this particular conception.

But a narrow conception of republican liberty is another matter. If limited to the prohibition of arbitrary treatment, republican liberty is useful – indeed, essential – for instantiating a great many comprehensive conceptions of positive liberty. It can therefore be seen as part of what John Rawls calls a ‘thin conception of the good’, something which all reasonable people can embrace, and something that a liberal government can promote without violating either the principle of neutrality or the principle of toleration.⁶⁹ And unlike thicker conceptions of republican liberty or express conceptions of positive liberty, it would not conflict with other rights. After all, no one worth taking seriously argues that under liberalism, some people have a right to treat others as a slave. Accordingly, sourcing the right to free expression in republican liberty makes the connection between free speech and freedom but avoids Dworkin’s worry because it does not make the liberal conception of what it means to have a right impossible to fulfil.⁷⁰

2.10. Arbitrary treatment defined

Of course, a restriction can be arbitrary in a number of different ways, so we must have a little better idea of what this might mean in order to give this idea a basis for practical application. Under my conception, a decision or action is arbitrary only when it is based on no criteria; is made without relevant criteria being adequately considered; is based in whole or in part on criteria that are unrelated to the performance of the task or role under consideration; or is based on performance-related criteria but does not treat similarly situated people similarly. In other words, under my conception of republican liberty, arbitrariness can be the result of either a defect in the procedure that produced a particular outcome, or a defect in the nature of what has been considered and/or the weight assigned to various decision factors.⁷¹ For example, absent some very unusual circumstances, it would be arbitrary for the government to ban the public use of the word ‘when’. This is a trivial example, but more significant real-world examples would include whether a state can ban the display of the gay pride flag in schools,⁷² or, even more broadly, ban schoolroom discussion of sexual or gender orientation, a ban instantiated by what its opponents have nicknamed the ‘Don’t Say Gay’ law.⁷³ I recognise that whether these latter bans are arbitrary is more controversial than whether a ban on the word ‘when’ is, but examining whether a regulation is arbitrary in one of the senses I have listed is precisely the debate we should be having in deciding whether such a ban is a violation of free speech or not. Few principles generate clear answers as to what is within their scope in every case. It is nevertheless important to be arguing about the right things in resolving the question. And unlike many other standards, we have thousands if not tens of thousands of legal and administrative decisions going back more than a century giving examples of what is and is not arbitrary.⁷⁴ So even though the edge of arbitrariness, like many other distinctions, can be blurry, this does not mean the arbitrariness test is not still both important and very useful.⁷⁵

While any arbitrary government restriction on speech would be a violation of republican liberty, the main focus of this paper is on whether we can prohibit hate speech and disinformation and if so, how that kind of speech might be best defined. It will therefore be more important here to focus not on whether a prohibition on speech is arbitrary and therefore a violation of republican liberty, but on the counterpart of this question. And this is whether certain kinds of speech are violations of republican liberty themselves and not within the scope of protection of any right to free speech derived therefrom. For despite the vast amount of literature on the right to free speech, much of it – perhaps even most, with the exception of arguments against government reluctance to prohibit the publication of pornography – focusses on when government restrictions violate the right to free speech, rather than on the extent to which the government suppression of certain kinds of speech is necessary in a liberal society to protect the republican liberty of its members.⁷⁶ Both questions, however, are equally important. To have a functioning liberal society, we need to know not only when to restrain the government, we also need to know when to encourage the government to intervene. And recognising republican liberty as the source of the right of free speech helps us do that, and not just when applied to the question of pornography.

I will get to how we tell if a particular form of speech itself constitutes arbitrary treatment in a moment, but first I want to emphasise that there is an even more important category of speech that sourcing the right to free speech in republican liberty allows us to fill. And this, in turn, is what allows us to avoid some of the definitional problems and fine line-drawing that afflicts other attempts to deal with hateful, offensive and misleading speech. For not only is speech that amounts to the arbitrary treatment of others not within the scope of the right to free speech, speech that makes the arbitrary treatment of others more likely (I shall discuss how much more likely in a moment), is also not within the scope of that right.⁷⁷ For even libertarians – that is, liberals who place the utmost emphasis on the protection of liberty – recognise that ‘risky behavior’ – that is, behaviour that makes a rights violation more likely – can be a rights violation itself and prohibited by the state.⁷⁸ Indeed, a great many rights are rights against this kind of behaviour. Reckless driving, failure to insure, failure to comply with the applicable building code, practicing law or medicine without a license, are all forms of risky behaviour and may be prohibited and justly punished even if the violator does not cause any personal injury or property damage (note also that practicing law or medicine without a license is often instantiated through speech). Making a product that is unreasonably dangerous is also unacceptably risky behaviour, even if the product does not ultimately cause injury to anyone. The whole basis of holding all members of conspiracy criminally liable for all the wrongful acts in furtherance of the conspiracy committed by any member, even members who did not know of these acts in advance and would not have approved of them if they had, is based on the idea that joining a conspiracy makes these rights violations unacceptably more likely.⁷⁹ Selling and even possessing certain dangerous drugs is a rights violation even if they are not consumed and, if they were consumed, no one was injured, because of the risk of facilitating consumption that such behaviour nevertheless entails. Using certain means and methods of attack in warfare is a rights violation if it puts civilians at unreasonable risk even though no civilians are actually injured thereby.⁸⁰ And so on.

The degree of risk imposed matters in all these cases, of course, and when determining whether the risk is sufficient to constitute a rights violation in itself, we typically consider: the potential harm if the risk comes about; the probability that this harm will occur; the difficulty of limiting that risk and bringing it into acceptable limits; the social utility of the activity in question; the degree of confidence we have in our various calculations, and so on. The more serious the harm threatened, the less the degree of risk matters. The greater the degree of risk, the more likely we will want to limit that risk in some way regardless of the harm threatened. The greater the social utility of the activity in question, the more risk we may want to tolerate. There are many possible permutations here. Which means there is no set formula for these calculations – each represents a determination of what is reasonable under the circumstances. So there can be disagreements on the borderlines here. But as I shall discuss later when I talk about the distinction between the reasonable and the unreasonable, we have a great deal of experience in drawing this border around a wide variety of activities every day.⁸¹ And borderline cases are not what concern us here – with regard to speech, it is those that are clearly unreasonable and growing in frequency and volume. In any event, the point I am making here is simply that speech can constitute an infringement of republican liberty not only when it is an act of arbitrary treatment, but also when it imposes a sufficiently serious risk of arbitrary treatment. Not because we need to look to some other principle of liberalism to provide a limit on free speech, but because the imposition of such a risk is not within the scope of liberty from which the right to free speech is derived.⁸²

2.11. What about autonomy?

In some ways, my sourcing of the right to free speech is somewhat similar to what T. M. Scanlon was trying to do when he claimed that the right to free expression is rooted in autonomy.⁸³ But autonomy is itself a highly contested concept – its exact scope and precise application remains unclear, as does its status as a right. Like republican liberty, autonomy is usually seen as an aspect of positive liberty. True, like my narrow conception of republican liberty, it is also seen as less than a full theory of positive liberty. But it also seems to encompass ideas beyond those expressed by my narrow conception of republican liberty as the right to be free of arbitrary treatment. One is arguably not acting autonomously, for example, if one is acting pursuant to the influence of personal demons or some sort of false consciousness rather than the imposition of the arbitrary will of another human agent. The concept of autonomy therefore brings with it many of the elements that liberals criticise about more comprehensive theories of positive liberty – it suggests that other people may know what you need to do to be your true autonomous self better than you do yourself, and they therefore may be entitled to coerce you to be free.⁸⁴ And absent government coercion to enforce its particular view of autonomy, individual conceptions of autonomy can conflict.⁸⁵ So to the extent that Scanlon is indeed trying to refer to the same thing as I am here when he is talking about autonomy, I do not see any advantage to using Scanlon’s terminology. To the extent he sees autonomy as something other than fully encompassed by my narrow idea of republican liberty, he is not clearly sourcing the right to free speech to a kind of liberty that can be the basis of a right in a liberal society – indeed, the closer his concept gets to a broader conception of positive liberty, the more unlikely it is that it can be the focus of an overlapping consensus.⁸⁶ And in either case, the lack of clarity as to what exactly Scanlon is referring to and where it comes from makes it a less attractive substitute for the approach I am arguing for here.

The same can be said of C. Edwin Baker’s conception of autonomy in his ‘Autonomy and Free Speech’.⁸⁷ An additional concern here is whether Baker actually has a conception of autonomy – he expressly states that he sees autonomy and liberty ‘as largely interchangeable’, and when he is supposedly talking about autonomy, what he really seems to be talking about is what other theorists would commonly call negative liberty (he expressly rejects theories of autonomy that could be categorised as theories of positive liberty and self-realisation, which are more often associated with theories of autonomy). When he presented the same theory of free speech in his earlier Human Liberty and Freedom of Speech, in fact, he expressly claimed he was presenting a liberty-based theory,⁸⁸ and his discussion of liberty there again made clear that it was negative liberty he had in mind. And if Baker’s view is really about maximising negative liberty, not autonomy, then his view has some serious and well-known problems.

First, as Hillel Steiner points out in his Essay on Rights, whenever we increase the negative liberty of one person, we usually do so by imposing restrictions on the negative liberty of another.⁸⁹ And this is often the case with speech, given the impact that certain kinds of speech can have on the liberty of others to enjoy their lives. So it is not even clear that total negative liberty can be increased or decreased rather than simply redistributed.⁹⁰ If it cannot, or even often cannot, then the maximisation imperative it is not a coherent objective.

Second, how do we maximise negative liberty without making qualitative rather than exclusively quantitative judgments? The problems with taking a purely quantitative approach are well-known – as Charles Taylor pointed out many years ago, if we try to rely only on quantitative judgments, we end of thinking that Albania before the fall of the communism was more free than England because Albania had less traffic lights, even though no one in Albania had the liberty to practice their religion or criticise the government, and this is absurd.⁹¹ Surely when it comes to the liberty to speak, it is not the loudest or most frequent voices we want to protect over all others; what a speaker has to say matters in deciding this. And if we do make qualitative judgments (that is, compare and rank the different kinds of speech involved), then what theory is being used to do this? Baker does not say, and whatever theory this might be is then the one really doing the work, so it is misleading to think that a concern for negative liberty is really motivating anything here.

Finally, there are the powerful concerns that Dworkin raises about negative liberty-based theories of free speech and which I have already discussed. Baker does not even mention much less address Dworkin’s concerns in either of his works. So we could end our discussion of Baker’s approach here.

But some theorists do characterise Baker as presenting an autonomy-based theory,⁹² and he does mention autonomy quite a lot not only in his more recent article but in his earlier book, so let’s ignore how his argument is best characterised and simply take his argument as based on autonomy. And at times, his discussion of autonomy does indeed seem to equate a respect for autonomy with something like refraining from treating another arbitrarily. But he also suggests that any speech which is ‘coercive’ violates another’s autonomy. This seems to be a spillover from the importance he places on negative liberty, for interferences with negative liberty are also often said to be wrongful if they are coercive. But coercion is a much bigger category than arbitrariness, for coercion can be permissible or impermissible, and this ignores the difficulty of distinguishing coercive speech from speech that is merely ‘influential’, with or without the creation and exploitation of some background ‘false consciousness’ or other conditions that limit the choices people can reasonably be expected to make. We can, for example, entice another to enter into a contract by offering them something they would rather have (like money) than what they currently have (like a piece of property). But we cannot do so by holding a gun to their head. Determining what is merely permissibly enticing and what is impermissibly coercive requires the application of a pre-existing system of rights that tell us what we can and cannot do to one another.⁹³ But it would be strange indeed to think that the scope of the right to free speech, which most people view as one of the most important rights we have, could only be determined after we had decided on the scope of all our other rights.

There are other problems with Baker’s conception of autonomy as well. Given his focus on whether and to what extent speech limits the choices available to the hearer and therefore supposedly infringes his autotomy, what do we do when speech is made to a group of people and some are coerced by it, some are not, some are content to be influenced by it in ways that others are not, and so on? Is speech which coerces another impermissible no matter how idiosyncratic this other person’s view of autonomy is? Is it permissible as long as there is someone who is not coerced by it? Once again, I do not see how Baker’s conception of autonomy, if it is indeed possible to say that this is what Baker’s is relying on, has any fewer problems defining the scope of the right to free speech than the marketplace of ideas and democracy theories he attacks, and it seems to have a great many more problems than the approach I propose. So I will proceed on the basis that the best way to source the right to free expression is to see it as an extension of a narrow conception of republican liberty.

3. What does sourcing free speech in freedom in the form of republican liberty tell us about the scope of the right?

So if freedom of speech is derived primarily from our narrow conception of republican liberty under liberalism, what does this tell us about the right? The first thing to note is that if the right to free expression is derived from this fundamental conception of freedom, it cannot be broader than the fundamental principle of freedom itself. Which does not mean that the right to free speech cannot be limited by other fundamental principles – it can indeed be limited in ways that it would not be if those other fundamental principles did not exist, because fundamental principles, being general concepts that can be further refined in a variety of ways and not more precisely defined conceptions, can conflict depending how they are further specified. But rights cannot, at least once their scope has been properly defined, or so I have argued extensively elsewhere.⁹⁴ So the right to free speech may be shaped or limited by other fundamental principles, but it cannot draw added scope from anything other than freedom itself, for then we would again be venturing into the realm of the ad hoc.

One immediate advantage of seeing the right to free speech as sourced in republican liberty is that this allows us to make sense of a variety of common restrictions on free speech that many liberals already accept, but have still been unable to persuasively define much less explain and defend.⁹⁵ For example, using my approach, we can explain why ‘hate speech’ is not protected under liberalism and indeed should be suppressed. For hate speech is not an exercise of freedom but an attempt to infringe the freedom of others in the republican sense by making false and misleading claims and by using character assassination to subject certain groups and sometimes specific individuals to arbitrary and therefore unreasonable restrictions, and sometimes even physical violence. Other theorists often source the prohibition of hate speech to its infringement of the right to equality held by its victims.⁹⁶ But even if this represents a correct assessment of the right to equality (and this is controversial), sourcing the right in liberty rather than equality makes the prohibition of hate speech much more secure. For almost all liberal political philosophers think that liberty has priority over equality. Indeed, John Rawls – the preeminent liberal political philosopher of the twentieth century – contends that protecting a basic liberty like free speech from infringement has lexical priority over addressing inequality, meaning that no amount of basic liberty, no matter how small, may be traded off for an improvement in inequality, no matter how large.⁹⁷ This means that if one did have the liberty to engage in certain kinds of hate speech, equality would not justify any restrictions on such speech for it could not justify an infringement of liberty. So it is important that we show, as I have argued here, that prohibiting hate speech does not present a conflict between equality and liberty but is instead an expression of both.

Some hate speech, however, is purportedly based on sincerely held religious beliefs. These speakers and the theorists who back them up then argue that religious people have the religious liberty to engage in hate speech regardless of its impact on the republican liberty of the targets of that speech, especially with regard to hate speech directed at the LGBTQ + community.⁹⁸ Here we supposedly face a conflict not between liberty and equality, but between one form of liberty and another. But as I have pointed out at length elsewhere, no such conflict actually exists.⁹⁹ Religious liberty is the product of the liberal principles of toleration and neutrality – it is a liberty, not a right in the Hohfeldian sense.¹⁰⁰ This means the government cannot prohibit the pursuance of a reasonable religious way of life because it disfavours that particular religious view. It does not provide a basis for refusing to abide by prescriptions on speech or behaviours when these prescriptions are otherwise justified infringements of negative liberty, or when the religious view is unreasonable (I shall discuss the importance of the reasonable v. unreasonable distinction further in the next section). For example, human sacrifice is not permissible no matter how sincerely it is held as a religious practice, nor is speech calling for it. Nor is any other kind of hate speech that threatens the republican liberty of those subjected to it.¹⁰¹

Another advantage of seeing the scope of free speech as being sourced in republican liberty is that this explains why telling even knowingly blatant lies is sometimes consistent with liberal values, and therefore not something that should be suppressed at the time or punished after the fact. For example, lying to the Nazi at the door about whether there are Jews in your basement, or to the sheriff on a country road about whether there are escaped slaves under the hay piled in the back of your wagon, is not outside the protections of free speech in a liberal society. This is because neither lie threatens the republican liberty of those asking or those in hiding. On the contrary, these lies are an attempt to protect the republican liberty of those in hiding and they do not infringe the republican liberty of either interrogator. On the other hand, lying to a police officer about whether a white supremacist who is on the run after committing a mass shooting is hiding in your basement, is a different matter. In this case, the lie threatens the republican liberty of others, and not the republican liberty of the white supremacist in the basement or any other right he might possess.

Of course, there are other arguments for how such deliberate lies can be justified.¹⁰² To the extent these provide supplemental reasons for sorting out these cases rather than contradictory ones, I express no view on whether they are correct. I am simply arguing here that we can get to the same place by relying on a narrow conception of republican liberty as setting forth the scope of the right to free speech.

My suggestion that we use the kind of arbitrariness that is an infringement of republican liberty as the test for determining the scope of the limits on speech within a liberal society has yet another advantage: it allows us to avoid the difficulty of otherwise coming to an agreement on exactly what hate speech is, for the definition of this has become notoriously controversial.¹⁰³ While hate speech is almost always defined as disparaging a certain race, religion, or other identifiable group of people, similar to the old tort of group libel (a tort is a civil as opposed to a criminal wrong), other requirements are also often added into the definition. For example, some argue that in order to constitute hate speech, the speech must be ‘calculated’ to cause embarrassment or injury to the group being referred to, and sometimes, that the speech in question must be intended to incite violence or, even more strongly, imminent violence, against that group to be objectionable.¹⁰⁴ Certainly an incitement of violence is something we need to act quickly to prevent. But just because direct incitement is missing, or the violence incited is not imminent, does not mean we should tolerate such speech. After all, remember that Julius Streicher, the editor of Der Stürmer during the Nazi period, was convicted of crimes against humanity at Nuremberg for the cumulative effects of 25 years of infecting ‘the German mind with the virus of anti-Semitism’ and thereby indirectly contributing to the Final Solution.¹⁰⁵ And a consistent relationship has been found since again and again between violent rhetoric, even if general and indirect, and the rise of hate crimes and other forms of political violence against those subject to such verbal attacks.¹⁰⁶ Remember also that sourcing the right to free speech in republican liberty not only means that speech that constitutes an actual violation of republican liberty is not protected, it also means that speech that constitutes a threat of violation is not protected either. So even if the incitement of violence is indirect, it is not inconsistent with liberalism to see such speech as outside the scope of the right to free speech. How indirect it may be and still constitute a threat may remain fuzzy, but the answer to that will depend on who is speaking, the medium used and its reach, the context in which the speech is made, the danger it poses given the contemporary political climate, and so on. These may seem like difficult questions, but they are the kind of questions that are resolved by prosecutors, judges, and juries every day in the context of adjudicating whether something constitutes threatening behaviour in a wide variety of other circumstances.

This approach also has the advantage that it does not require us to tolerate speech that is not ‘intended’ to disparage a group even though it is reasonably clear that it will have this effect. Suggesting, for example, that the world is run by a secret cabal of 12 foot-tall child-sacrificing bloodthirsty alien lizard people, most of whom are disguised as Jews, as the former British footballer David Icke does, may be a genuinely held belief and, in some sense, it may not even be ‘calculated’ to injure Jews – Icke’s intent may be simply to inform the public of the ‘danger’ in their midst and not, as he says, to impugn ‘real’ Jews.¹⁰⁷ But if the only tool we have available prohibits hate speech that is intended to and actually does qualify as an imminent incitement of violence, some outrageously false and historically dangerous statements are going to remain within the protection of the right to free speech when under the ‘encouraging arbitrary treatment’ test they would not.¹⁰⁸

But we want more from our approach to delineating the scope and limits of free speech in a liberal society than simply providing a good explanation for our current practice with regard to hate speech and solving some uncomfortable problems that arise under some of its would-be definitions. We want it to help us go beyond this and guide us toward the regulation of the kind of false and misleading statements that are so damaging to the fabric of social cooperation and democracy in the modern world, even when they do not constitute hate speech, broadly construed. And while direct attempts to incite hatred and discrimination and violence on arbitrary grounds are not within the scope of free expression given its source in republican liberty, attempts to disseminate information that is false or misleading will not always or perhaps even often fall into this category. In many cases, the causal relationship between such speech and the arbitrary treatment of others is less obvious and direct, which is how those on the far right have managed to get away with what is inflammatory speech despite the prohibition of hate speech, even when this prohibition happens to apply. After all, people often say things that are untrue, but this alone is not sufficient to make such statements outside the scope of a right to free speech that is sourced in republican liberty. To see where the line between protected untruths and protected ones is, we need to consider the limits imposed on liberty by the liberal principles of toleration and neutrality.

4. How we can derive limits on free speech from the very principles that limit freedom itself

4.1. Limits derived from toleration and neutrality

I have already discussed how principles related to the maintenance of a healthy democracy and to the discovery of truth can supply limits on the right to free speech, so I want to focus on something less obvious here. And this is the extent to which we can find limits on the right in the more general principles of toleration and neutrality. There is wide agreement, of course, that the principles of toleration and neutrality are among the important values being instantiated by the liberal conception of freedom. Indeed, an aversion to the principles of toleration and neutrality is one of the primary grounds upon which the illiberal right has challenged liberalism for more than a century.¹⁰⁹ But while toleration and neutrality are clearly part of what the liberal conception of freedom is designed to implement, these principles also tell us where some of the limits to freedom lie. For example, just as the principle of toleration does not require toleration of the intolerant, the advocacy of intolerance is not something that freedom and therefore free speech was intended to protect, even when this does not meet the technical requirements of hate speech, or so I shall argue in a moment. Similarly, while the principle of neutrality requires government to be neutral between competing conceptions of the good in a liberal society, this applies only to conceptions that are reasonable – the government need not be neutral when it comes to conceptions that are unreasonable.¹¹⁰ And the same distinction must be made between the reasonable and unreasonable when it comes to speech too. Speech that is unreasonable – to be defined further in a moment – is accordingly not something toward which the government must be neutral.

I know, many liberals will want to resist the claim that toleration and neutrality can be the source of limits on free speech as well as supportive of its protections. Following First Amendment jurisprudence, they want to say that with regard to speech, these principles prohibit any form of ‘content-based’ regulation, no matter how offensive that content might be. And there are of course some free speech absolutists who contend exactly this. But most of those who embrace the prohibition on content-based restrictions are uncomfortable with what this would mean when taken to the extreme, and therefore are willing to engage in even strenuous intellectual contortions to carve out exceptions to this – to explain why a prohibition on lying is not content-based, for example, and why the prohibition of hate speech is based not on the ideas expressed, but on the actions incited (and the same with pornography).¹¹¹ But this two-steps-forward one-step-back manoeuver is born out of a gross misunderstanding of what the liberal principles of toleration and neutrality are there to do. These principles are in fact not there to prohibit government from regulating on the basis of an objection to the substance or nature of an act, whether it constitutes speech or otherwise.

What the principle of toleration is there to do is to ensure reasonable pluralism; it is not there to protect the expression and pursuance of all conceptions of the good no matter how unreasonable a particular conception might be. Under the principle of toleration, it is therefore not only unreasonable to advocate intolerance of the reasonable; it is also reasonable to advocate intolerance of the unreasonable. And if it is reasonable to advocate intolerance of the unreasonable, it is reasonable to do something about this too; that is, to regulate intolerant speech in an appropriate manner. ‘Big tentism’, the colloquial way of referring to the embrace of reasonable pluralism, does not require that we invite into the tent people who would throw everybody else out. And not just because tolerance of intolerance is harmful, both to the those subjected to it and to society at large, although of course it is.¹¹² There is also a deontological reason for such regulation. For as Rawls himself pointed out long ago, the intolerant have no claim in justice against those who would suppress them, because they would do the same if they were in a position of power.¹¹³

Similarly, the principle of neutrality is there to ensure that the government does not choose between the various reasonable conceptions of the good that a pluralistic people happen to embrace. But this does not imply that the government may not choose between reasonable and unreasonable conceptions of the good. I have defended this claim elsewhere at length,¹¹⁴ so I will not say more about it here, but I will note here that if this were not true, then it would be impossible to recognise rights, for rights are based on substantive objections to the conception of the good that the violator is furthering. For example, the right against being murdered is not value neutral; it reflects a content-based judgment that murder constitutes unreasonable conduct.¹¹⁵ The same is true of every other substantive right I can think of. (It is possible that procedural rights might be construed as value neutral – but the point is not whether a right could be value neutral, but whether we could have the system of rights we actually have, many of which are substantive, while still remaining value neutral.) So when it comes to speech, the principle of neutrality does not prohibit the consideration of content-based factors – it merely prohibits the government from infringing on speech when the content of that speech is reasonable rather than unreasonable, a distinction I shall say more about in a minute.

Of course, it may be that the fear of recognising fetters on speech may not come from concerns about the requirements of toleration and neutrality but from purely practical considerations. That is, many liberals no doubt fear that tighter limits on free speech will be used first and primarily to stifle the kind of voices they most want to protect.¹¹⁶ And it is indeed true that authoritarian regimes or regimes with authoritarian tendencies, which I think fairly describes the recent Trump administration and will the next one if Trump wins another term,¹¹⁷ often abuse whatever limits are accepted on speech to silence their critics.¹¹⁸ This is also true for the mini-me Trump administrations in power in various US states and in other nations, and will no doubt continue to be true under the administration of any like-minded authoritarians currently waiting in the political wings.¹¹⁹ But all anti-liberal regimes that rise to power within the liberal world, or even ostensibly liberal regimes that merely have authoritarian tendencies, will engage in such attempted suppression anyway. They will merely find some other principle to distort and misuse – national security is one that always seems to be available here – or dispense with the worry of trying to seem principled whatsoever. And this means that liberals who are afraid of recognising limits on freedom of speech because of practical concerns are misjudging the degree to which this will hamper the enemies of liberalism in their pursuit of what are fundamentally illiberal goals and end up handicapping the defense of liberalism instead.

Some liberals also think that we must tolerate the intolerant given the liberal commitment to settling disputes by compromise and negotiation and as a way of showing what a powerful idea liberalism is, so powerful it need not fear advocacy of its enemies’ position. But this view seems driven more by arrogance and ego than by principle, and so makes a poor reason for withdrawing from the battle for ideas. No doubt, some of these people also want to tolerate the intolerant because as the enemies of liberalism often claim, liberals can be overly-averse to conflict and reluctant to fight for what they believe. Hence the popular quip by the conservative poet Robert Frost – 'a liberal is someone who cannot take their own side in an argument’.¹²⁰ This latter reluctance becomes even stronger, of course, when it seems like the intolerant ideas at issue are so self-evidently crazy that they will be unappealing to the masses, and therefore can be comfortably viewed by liberals as non-threatening.

But while this ‘softly-softly’ approach is not problematic when unreasonable claims are only made by those on the fringe and therefore pose no real threat to our social life in general or to any group or individual, this is not an accurate way of characterising the threat that liberal societies now face. Until relatively recently, a speaker could only reach an enormous audience through mediated forms of communication. These mediators have always been private parties, it’s true; but they played a critical role in making liberal democracy work. The current private parties who are supposed to perform this role, however, have been allowed to opt out and provide essentially unregulated platforms to whomever wants to use one.¹²¹ It is accordingly now possible for anyone to address the entire world without their views being mediated by anyone. As a result, the dissemination of false and misleading information and the advocacy of intolerance and other unreasonable ideas has become so widespread that it now threatens the continued existence of liberalism. So liberal grace – the willingness not to enforce the requirements of liberalism, is long past its sell-by date.

4.2. Other ways that the distinction between the reasonable and the unreasonable can help us here

Besides showing that we have no obligation to protect the advocacy of intolerance and unreasonable ideas, however, can we say more about what the distinction between the reasonable and unreasonable can help us with here? To do so, I will employ the concept of the burdens of judgment, which Rawls uses to explain why reasonable people can reasonably disagree.¹²² Evidence, for example, is often complex, conflicting, and hard to assess, as is the weight to be assigned to relevant considerations; normative concepts are often vague and open to different interpretations; and our judgments on these matters are influenced by our personal experiences and our personal experiences are diverse, leading us to reasonably assign different priorities to competing considerations. Because of the burdens of judgment, it is not unreasonable to expect good faith disagreement about the scope or limits of the principles of liberalism or their application. Most importantly, while we may be convinced that the content of certain speech is mistaken and misleading and even dangerous in important ways, we should nevertheless tolerate some degree of error as a concession to the difficulty involved in coming to view about these matters. Applied with these ideas in mind, then, the concept of the burdens of judgement provides a tool we can use to help us implement the distinction between the reasonable and the unreasonable here. It can help us discriminate between statements that are false or misleading but reasonable, and those that are so implausible and unsupported by any evidence or reasonable inferences to be made therefrom, that no reasonable person could believe they were true. And in the latter case, this tells us that such statements are not entitled to protection in a liberal society, one that is dedicated to the idea that reasons for action should be based on the most reliable information derived from a system designed to ensure widespread access to such information.

Note that by using the burdens of judgment test as a way of applying the distinction between the reasonable and unreasonable to speech, we can also avoid coming to the absurd position some liberals adopt to the effect that if a significant enough portion of the public believes something to be true, then under liberalism, the government has an obligation not to suppress that view, and private parties may even have an obligation to provide a platform for those advocating that view, regardless of its established falsity.¹²³ But reasonableness is an objective standard – under negligence law, for example, conduct may be found unreasonable even if the majority of people do it.¹²⁴ And if we apply this same idea to the burdens of judgment test, counting how many people believe something to be true is not evidence that it is reasonable to believe it is true if that belief is unfounded. Besides, the portion of the population that believe a falsity is often if not always the result of this falsity being promoted by people who have political or social influence. The percentage of the population who believe such a lie is therefore an effect of not regulating speech that is beyond the burdens of judgment in the first place – the effect of this failure cannot then be used to justify including some radically false belief in the burdens of judgment.

Admittedly, in fringe cases, it may be hard to tell where to draw the line between cases that are just within the burdens of judgment and those that are just outside it. But fringe cases are not the problem here – the problem is that we currently lack a rationale for dealing with clear cases. And the burdens of judgment test solves this problem. For example, claiming that Hilary Clinton is running a pedophile ring out of the basement of a Washington pizzeria is simply not a claim that is even remotely within the burdens of judgment.¹²⁵ Attempts to spread such bizarre and yet dangerous falsehoods are accordingly not protected by the right to free speech, and therefore may be suppressed within the limits of liberalism for the normal kinds of reasons justifying any kind of regulation. And these kinds of non-protected statements should be suppressed, for if they are not, there is a risk of significant harm, a risk that is amply demonstrated by the invasion of the pizzeria in question by an armed vigilante seeking to ‘rescue’ the children supposedly being held by the ‘ring’ in the pizzeria’s ‘basement’, which did not even exist.¹²⁶ After all, using force, even deadly force, is not a crime when it is reasonably believed to be necessary for the defense of others, even if this belief turns out to be mistaken. So unless we are prepared to find that the ‘pizzagate’ shooter committed no crime, then we must be prepared to treat the claim of what was going on there as beyond the burdens of judgment and therefore as unreasonable and unlawful.

There are of course many, many other claims I could mention here as being uncontroversially outside of the burdens of judgment and therefore not something to which liberalism gives free speech protection. The claim that the shooting at Sandy Hook Elementary School did not occur and that the apparently distraught parents were really ‘crisis actors’ hired as part of a vast conspiracy to undermine gun rights, a claim which was repeatedly aired by right-wing Infowars radio host Alex Jones and others and which, not surprisingly, inspired a number of individuals to engage in the brutal harassment of many of the grieving parents and contributed to the suicide of one, is obviously outside the burdens of judgment.¹²⁷ The continuing claim that climate change is a hoax, or that it is not being driven by human activity, despite the overwhelming amount of evidence and almost universal scientific consensus to the contrary – lies that are being pushed by the oil industry and right-wing billionaires¹²⁸ in an effort to protect their fortunes – are outside the burdens of judgment¹²⁹ and seriously undermining our ability to make the sacrifices necessary to maintain our physical environment.¹³⁰ The claim that vaccinating children puts them at increased risk of autism is similarly false and misleading, and unreasonably and dangerously so, because it puts all of us (and not just the children who remain unvaccinated) at risk for contracting diseases to which we need not be exposed.¹³¹

One of the biggest and most destructive lies currently ripping holes in the existing social fabric of liberalism, however, is the continuing claim by Trump and his supporters that he won the 2020 election ‘by a lot’ even though he lost it by a staggering 7 million popular votes and 74 electoral votes, and that the election was accordingly stolen from him by a vast conspiracy of Democrats, the liberal media, the conservative Fox News, numerous Republican governors and secretaries of state, an assortment of Venezuelan agents including the deceased former President of Venezuela Hugo Chavez, and a variety of other suspects both usual and unusual. Initially, this lie was repeated and promoted freely on social media. Eventually, it was accompanied on some sites by a warning that the claim was ‘disputed’, or simply by a hyperlink referring to a site where contrary information could be found (for example, ‘see election results’).¹³² But this warning is woefully insufficient to convey the true character of the claim. For the latter kind of reference does nothing to tell people who do not click on it that the post they are looking at contains false or misleading information. Even the warning ‘disputed’ can easily be understood to mean ‘disputed in good faith’, and therefore as indicating that there is conflicting credible evidence and reasonable inferences to be made therefrom on both sides. This, in turn, falsely suggests that each of the opposing claims here are within the burdens of judgment.

But the claim that the election was stolen was actually supported by no credible evidence.¹³³ And while absence of evidence is not evidence of absence, in this case, there is plenty of evidence of absence as well. Indeed, there is overwhelming evidence that voting in this case was remarkably free of fraud or hacking.¹³⁴ In other words, there is overwhelming evidence that the claim that the election was marred by widespread fraud and other unlawful behaviour is not within the burdens of judgment. Merely marking this claim as disputed is accordingly misleading in its own way and therefore not a way for social media companies to behave responsibility. Only removing such a claim in its entirety and preventing people from repeating it, then barring those on the platform who attempt to do so anyway at first temporarily and then permanently if they persist, can prevent this kind of lie from undermining trust in our basic institutions, trust that is essential for our continued enjoyment of republican liberty to persist. And nothing in the liberal conception of toleration or neutrality, or in a republican liberty-derived right to free speech, says otherwise.

4.3. But what about the difficulty of line-drawing in less outrageous cases?

Unfortunately, no matter how strongly we might wish it were otherwise, no political theory is judgment-proof. That is, no theory worthy of the commitment of serious people provides a fully-determinate easy-to-apply standard for enforcing its beliefs. Only a theory that said everything was permitted or that nothing was permitted could do this. As a result, the limits applicable under every theory worthy of our embrace are subject to abuse. For no matter where the line is, there will always be cases that straddle it. We can reduce the importance of these cases, however, and reduce our fear of getting too close to the edge of the slippery slope of unjustified suppression, by recognising that we should only enforce the line in clear cases, as we do when imposing criminal sanctions. If we operate in this way, we ensure that suppression is legitimate under liberalism only in a small category of cases. Denying that this category exists may make cases of improper suppression rarer (although I doubt this), but it also prevents liberals from standing up for what they believe. Yet liberalism, like the various constitutions that serve to instantiate it, is not a suicide pact.¹³⁵ It does not require that we stand passively by while illiberal forces work toward the destruction of the open society which liberalism has built. Liberalism has teeth as well as welcoming arms, and sometimes the principles of liberalism require that we use them.

Of course, this does not mean liberalism allows us to disregard people’s rights whenever we face what our leaders perceive as emergency conditions, as many on the far right ironically claim we should. It means that we must respect people’s rights no matter what conditions we may find ourselves in. But it also means that in defining these rights we recognise that there are borders on their scope and, even within this space, there are limits on these rights created by the necessity to harmonise all our rights into a coherent set. When people exceed the scope of their rights or transgress the limits that apply to them, we should be prepared to take appropriately proportional action in response.¹³⁶ The difference between authoritarianism and liberalism is not that one suppresses everyone who deviates from the one true path and the other suppresses no one no matter what they do. The difference is that one requires total unthinking obedience to the official government point of view and vigorously and disproportionately punishes not only any deviations no matter how small but also all advocacy of non-compliance. The other recognises suppression as only appropriate in a pre-defined narrow range of cases that go beyond the bounds of reasonable pluralism and even then, only by using the least-restrictive sufficient means.

Those who are concerned that the enforcement of existing limits on toleration and neutrality to regulate political speech will merely relocate noxious anti-liberal activity to even less responsible platforms and not reduce it should note that there is at least some evidence that this is not true.¹³⁷ Eliminating spaces on the internet where those engaged in hate speech meet does seem to actually reduce the incidence of that speech, rather than merely drive it to some other location.¹³⁸ The banning of Milo Yiannopoulus, the noxious alt-right provocateur, and Richard Spencer, the American Neo-Nazi, and other alt-right speakers from various social media platforms was not intended to, nor did it actually prevent, the ideas these individuals espouse from getting out. But this did serve as an important public condemnation of their views, a condemnations that far too many people today needed to hear. So there is something to be gained in the defense of liberalism here, and not merely, as some liberals seem to think, something to be lost.

Indeed, even a small number of misleaders can have a devastating effect.¹³⁹ Trump is a serial liar and misleader of epic proportions.¹⁴⁰ When he was finally banned from mainstream social media platforms for promoting lies about the 2020 election after these lies provoked violence on January 6,¹⁴¹ the effectiveness of this ban was astonishing – misinformation about the election immediately plunged 73 percent.¹⁴² Similarly, misinformation spread by anti-vaxxers is largely spread by just a handful of people (in this case only twelve).¹⁴³ No widespread purge is accordingly necessary here; identifying and banning super-spreaders of hate and disinformation, or deplatforming them as it is commonly called, is an effective even if not complete remedy.¹⁴⁴

Of course, sometimes the number of misleaders involved is larger, but the numbers are not so large as to be beyond the ability of social media companies and their sophisticated algorithms to control. At one time, twitter identified and banned more than 70,000 accounts for promoting one kind of disinformation or another (Covid-19 misinformation, election denialism, QAnon-based conspiracy theories, and so on). Again, more could have been done here, and done more quickly,¹⁴⁵ but the general effectiveness of these bans is demonstrated by the fact that the spread of this kind of disinformation rose significantly again after Elon Musk bought twitter and promptly rescinded the ban on just 1,000 of these people.¹⁴⁶ The same general albeit incomplete effectiveness can be seen when social media platforms tried to limit the spread of dangerously misleading disinformation about the coronavirus and its treatments during the pandemic.¹⁴⁷ So a consistent, more timely, and more responsible vetting of social media users is both possible and necessary.¹⁴⁸

4.4. Heading off some potential criticisms

Just so there is no danger of my being misunderstood here, let me make clear that I am not suggesting people should not be able to say things on social media, or anywhere else for that matter, that are false. Many false statements – indeed, most statements that turn out to be false, are nevertheless within the burdens of judgment. Rational people rarely make or seek to promote statements that are otherwise given the fear of not being taken seriously. But the proliferation of outrageous lies and misleading statements has proved that people are far more gullible than we may have thought. So more needs to be done to control this kind of speech. And the way to do this is to require social media companies to police their platforms, not just for incendiary statements – that is, those that even in their cautious view are likely to provoke violence, but also for statements that are clearly beyond the scope of the burdens of judgment, whether their incendiary character is obvious or not. If only these companies had taken action on this basis, as many had repeatedly urged them to do before the insurrection of January 6, that disgraceful event might not have occurred.¹⁴⁹ For while it is better for these companies to have taken action late as opposed to never, once these kinds of lies have been spread to the extent that they actually incite violence, they are going to be so embedded in the consciousness of so many that they are going to be difficult to dislodge even if these ideas are no longer permitted to be promoted. To protect liberalism, we simply have to be prepared to prevent claims that are beyond the burdens of judgment from being introduced into the fabric of our social life in the first place.

But is the cure I am suggesting here worse than the disease?¹⁵⁰ Here, the argument goes that it is inappropriate suppression that most often leads to or at least extends periods of tyranny, not the failure to employ suppression when it is appropriate to do so. And in liberal societies, there are certainly examples of inappropriate suppression, in most cases of left-wing voices, at least in recent memory. But we must not let fears that were appropriate in the past dominate a future where they are misplaced. It is true that our sense of the burdens of judgment change over time, and that even in liberal societies suppression has been inappropriate in the past. But in a liberalising world, our sense of the burdens of judgment is generally becoming broader, not narrower. The fear that it might be getting narrower again is the fear that the world is no longer liberalising. And if I am right, and the failure to suppress speech beyond the burdens of judgment is feeding this reversal in direction, then it is the fear of unjustified tolerance not the fear of justified intolerance that should be paramount for liberals.

Of course it is still possible that mistakes will be made. But it is still possible that some people will get convicted of crimes they did not commit, despite all our safeguards. This is nevertheless not seen as a reason to get rid of the criminal trial process, only to improve it. To ensure that the number of mistakes here are kept to a minimum, we can and should ensure that the restrictions imposed on outright suppression, the most serious remedy available here, are substantial, thereby minimising the threat of unintended damage. To be appropriate, suppression must be the least restrictive means that is likely to effective under the circumstances; it must be clear that the speech in question is beyond the burdens of judgment rather than a close call (in another words, a mere preponderance of the evidence, the usual evidentiary standard, is not enough); the danger from doing nothing must be high and the threatened damage severe; and even so, we must afford a reasonable opportunity for those banned to challenge any determination that has been made against them. If we respect those limits in good faith, the chances of a mistake should be minimal. And if we do move into an era characterised by a lack of good faith, then the suppression of reasonable speech will be an effect of that move, not a cause. Ensuring unfettered access to social media does not protect liberalism, it renders its aspirations impossible to achieve.¹⁵¹

4.5. A final word on the practicality of the ‘Unreasonableness’ standard

To those that are still sceptical that the test of ‘unreasonableness’ is determinate enough to be practical, remember that we actually have a great deal of experience in applying this same standard in other contexts. Lawyers cannot make statements in court or in legal briefs without a reasonable basis for believing that they are supported by the facts and the law; hence the disciplinary proceedings brought against a number of Trump’s lawyers for their role in trying to overturn the 2020 election.¹⁵² Indeed, the entire criminal justice system is designed to sort claims that are reasonable from those that are not – the reasonable doubt standard being the most prominent but not only example of this. The concept of reasonableness also appears throughout the civil justice system. Tort law prohibits a wide variety of unreasonable conduct, including negligence and, in some jurisdictions, even negligent misrepresentation.¹⁵³ Under contract law, a contract is created if one party creates the reasonable belief that the other party intended to be bound, regardless of their actual intent.¹⁵⁴ Under defamation law, the usual standard is that the false statement must have been made with a lack of reasonable care as to its truth or falsity.¹⁵⁵ Under strict products liability law, one cannot sell products that that are ‘unreasonably dangerous’. And so on. Taken together, these rules are generally seen as ensuring that the ‘free market’ for products functions properly and not as interfering with it, and such rules are generally credited with making society safer and producing a higher quality of life for everyone.¹⁵⁶ And they have generated a wealth of examples clarifying the reasonableness standard and therefore providing a basis for deciding what else is reasonable and what is not. So why should we be concerned that an unreasonableness standard, when applied to the right to free speech and the marketplace of ideas, is not practical? It is certainly more determinate and therefore more practical than more limited alternatives that seek to exclude only lying, which requires an examination of the subjective sincerity of the speaker,¹⁵⁷ something not readily available to proof in most cases and easy to falsify and disguise. And it does not leave those who make claims that are clearly beyond the burdens of judgment and therefore unreasonable even if sincerely felt, off the hook.

4.6. Reasons to question the distinction between commercial and political speech

Of course, many of the examples I have cited are cases that involve commercial or otherwise non-political speech, and many people believe that political speech is different. Fifty years ago, one prominent conservative even argued that political speech is the only kind of speech that should be protected by the right of free speech.¹⁵⁸ And that view has some contemporary adherents. Facebook, for example, has recently declared its intention to continue to refuse to police political speech, no matter how uncontroversially false or misleading it may be.¹⁵⁹ But why? Even political speech has always been subject to reasonable time, place, and manner restrictions. While what we are talking about here envisions going beyond this and regulating extreme speech on the basis of its content, it seems hard to see why it is permissible to police claims about diet pills and dessert toppings but not claims about matters that more directly impact the most fundamental arrangements under which we live. Not about whether they are false or not, but about whether they are so false it would be unreasonable for any person to think that they might be true. In fact, we already recognise limits of this nature on the content of speech in a wide variety of contexts. Many kinds of crimes or civil regulatory violations are also proved by reference to speech (price fixing, securities and other kinds of fraud, drug and other product mislabelling, and so on). Why should speech that justifies civil fines or enhanced penalties for certain criminal acts or constitutes a criminal act in and of itself not also justify regulation within the realm of the political? If prison sentences can be imposed or at least lengthened for speech in certain contexts, clearly lesser forms of regulation should also be permissible within the bounds of the concepts of toleration and neutrality.¹⁶⁰

I should note here that political speech – or at least speech that is directed at public figures – is currently within the American constitutional right of free speech unless such speech is made with ‘actual malice’. To be made with actual malice, in turn, speech must be made knowing that it is false or with reckless disregard to its truth or falsity, rather than merely negligently.¹⁶¹ Does this not conflict with my argument that the test here is one of reasonableness? No, because the proof requirements I incorporate effectively require the same thing. A statement that could not reasonably be believed to be within the burdens of judgment as proved by clear and convincing evidence rather than merely a preponderance of the evidence is effectively reckless. I do not have enough space remaining to explain the difference between these standards in further detail, but those who remain concerned about this can simply refer to the relevant legal textbooks. In any event, I do not see any significant conflict between what I am proposing liberal political morality requires and current American defamation law as moderated by the First Amendment.

In any case, the justification usually offered for protecting political speech more than other speech is that political speech is more important and therefore we need to be more careful when regulating it – it is, to quote those who make this distinction, ‘high speech’ as opposed to ‘low speech’.¹⁶² But I fail to see why this should make it more immune from regulation. Indeed, in any practical sense, the opposite is true. The damage caused by political speech that is outside the burdens of judgment is often going to be far more severe than that caused by non-political speech outside these burdens precisely because political speech is about more important and wide-ranging matters. The ‘Big Lie’ has brought us to the brink of the end of liberal democracy as we know it and may yet push us over the edge regardless of the result of the next election.¹⁶³ And the same is true of other ‘big lies’ being promoted elsewhere in the world too.¹⁶⁴ In other words, the justification offered for treating political speech more generously – that it is more important than other kinds of speech – also justifies treating it more restrictively. So I see no rational reason for not subjecting political speech to the same limits that other kinds of speech are subject to under liberalism.

Besides, distinguishing between speech that is political and speech that is not is extremely difficult to do, especially in the current, highly polarised climate.¹⁶⁵ Almost everything one could say today is political in some sense. Climate deniers and anti-vaxxers are arguably engaging in political speech and not just questioning well-established science. Even those who claimed Sandy Hook was a hoax were arguably engaged in political speech by tying what they claim ‘really’ happened there to gun rights. Anti-Semitic and racist and homophobic claims are clearly political, even if they are merely about the supposed cultural or scientific attributes or practices of the target group. Everything that Donald Trump says is arguably political, not only while he was president but now that he is an ex-president, simply because he said it. In a liberal society, one cannot falsely shout ‘Fire!’ in a theatre and cause a panic.¹⁶⁶ We must take note that due to advances in technology, we are now all technically sitting in the same theatre. A more proactive stance from government is accordingly required.¹⁶⁷

Unfortunately, things seem to be moving in the opposite direction – perfectly reasonable statements are getting suppressed while outrageous ones are being increasingly ignored.¹⁶⁸ And the rise of artificial-intelligence driven bots and ‘deep fakes’ is making matters even worse.¹⁶⁹ Alongside the failure of social media companies to adequately police the marketplace they provide to English speakers, almost nothing has been done to limit the spread of disinformation and conspiracy theories through social media aimed at immigrant communities, such as Asian and Latino communities in the United States.¹⁷⁰ Indeed, this goes some way toward explaining how Trump managed to improve his share of the votes within these communities despite his blatantly racist attitude toward them. Facebook has also allowed itself to be bullied into hiring more right-wing organisations as factcheckers because independent factcheckers find far more inaccuracies in statements coming from right wing sources – not because these independent fact-checkers were biased, but because in today’s climate the spread of fear, hatred, and disinformation is primarily a tactic of the right not the left.¹⁷¹ What we need is more vetting using the standards of liberalism to decide what threatens to incite arbitrary behaviour, what is within the burdens of judgment and what is not, and what constitutes unreasonable behaviour and what does not. And if we do that, the right to free speech can no longer be weaponised against liberalism but will remain a tool of maintaining it.

Acknowledgments

Earlier versions of this paper were presented at conferences in Budapest and Bucharest, both in 2023. Thanks to Julija Perhat for organising the first and to Paula Tomi for organising the second. And thanks to all those who participated for their comments and criticisms. Thanks also to Peter Colby, Harriet Davidson, and Lisa Herzog for much helpful discussion of the ideas presented here, and to one particularly thoughtful anonymous referee for this journal, whose numerous comments and suggestions led to many improvements in my argument.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 See, e.g., Steven Pinker, ‘Are College Campuses Infringing on First Amendment Rights? One Liberal Scholar Says Yes’ WBUR (24 January 2024); Michelle Goldberg, ‘Do Progressives Have a Free Speech Problem?’ The New York Times (17 July 2020) (citing left-libertarian Ellen Willis for proposition that ‘progressive movements sow the seeds of their own destruction when they become censorious’); Andrew Marantz, ‘Free Speech is Killing Us’ The New York Times (4 October 2019) (some speech might be bad, but censorship is always worse, say liberals and conservatives alike).

2 See generally Adam Liptak, ‘How Conservatives Weaponized the First Amendment’ The New York Times (30 June 2018) (discussing how those on both the left and the right become free speech absolutists only when it is convenient); David French, ‘The Moral Center is Fighting Back on Elite College Campuses’ The New York Times (16 April 2023) (same); Richard Stengel, ‘Why America Needs a Hate Speech Law’ The Washington Post (29 October 2019).

3 See Nick Robins-Early, ‘A Deranged Ploy’: How Republicans Are Fueling the Disinformation Wars’ The Guardian (10 July 2023).

4 See, e.g., Mark R Reiff, In the Name of Liberty: The Argument for Universal Unionization (Cambridge University Press, 2020) 230–238 (regarding ‘compelled speech’).

5 See, e.g., Owen M Fiss, Liberalism Divided: Freedom of Speech and the Many Uses of State Power (Westview Press, 1996); James Weinstein, ‘A Constitutional Roadmap to the Regulation of Campus Hate Speech’ (1992) 38 Wayne Law Review 163–247; Steven H Shiffrin, The First Amendment, Democracy, and Romance (Harvard University Press, 1990).

6 See, e.g., Janus v. AFSCME, where the Court used an aggressive reinterpretation of free speech doctrine to overturn decades of labor law jurisprudence allowing unions to charge non-members ‘agency fees’. For further discussion of this case and how it gets the right to free speech wrong, see Reiff, In the Name of Liberty (n 4) 230–238.

7 See, e.g., Reiff, In the Name of Liberty (n 4).

8 See, e.g., Mark R Reiff, ‘Neutrality and Excellence’ in Mark McBride and Visa AJ Kurki (eds), Jurisprudence Without Trimmings: Essays in Honour of Matthew Kramer (Oxford University Press, 2022) 271–96; Mark R Reiff, ‘Trump and the End of Liberalism: Some Hard Truths About the Degradation of American Democracy’ The Critique (15 January 2017); Mark R Reiff, ‘The Attack on Liberalism’, in Michael Freeman and Ross Harrison (ed.), Law and Philosophy (Oxford: Oxford University Press, 2007) 173–210.

9 See, e.g., Patrick J Deneen, Why Liberalism Failed (Yale University Press, 2018).

10 For further discussion of the difference between a concept and a conception, see Ronald Dworkin, Law’s Empire (Harvard University Press, 1986) 70–72.

11 See Reiff, ‘Neutrality and Excellence’ (n 8); Reiff, ‘Trump and the End of Liberalism’ (n 8); Reiff, ‘The Attack on Liberalism’ (n 8).

12 Inexplicably, one critic of this paper insisted that when I use the word ‘liberalism’, I am to be understood as referring to ‘classical liberalism’, the name some right libertarians use to describe themselves in an effort to re-brand their view as the original and therefore ‘truest’ form of liberalism. See, e.g., FA Hayek, The Constitution of Liberty (Routledge, 1960) 344–55; James M Buchanan, Why I, Too, Am Not a Conservative: The Normative Basis of Classical Liberalism (Edward Elgar, 2006) 52–61, 63. The critic then went on to contend that as a right libertarian theory of free speech, this paper is a failure because it argues for more government regulation of speech, not less. No political theorist of whom I am aware, however, including but not limited to those who call themselves ‘classical liberals,’ understands the word ‘liberalism’ and the term ‘classical liberalism’ to be synonymous; nor do any of them understand the term liberalism to exclude all otherwise liberal non-right libertarians. So the reading made by this critic would be odd even if I were I right libertarian and had said nothing about what I mean by liberalism in the text. But I do, and as should now be absolutely clear, I am not a right libertarian, I am not proposing a right libertarian theory of free speech, and a right libertarian definition of liberalism, whatever that might be, is not the definition of liberalism I am employing here.

13 See, e.g., James Weinstein, ‘Participatory Democracy as the Central Value of American Free Speech Doctrine’ (2011) 97 Virginia Law Review 491–514; Robert Post, ‘Participatory Democracy and Free Speech’ (2011) 97 Virginia Law Review 477–89; Martin H Redish, ‘The Value of Free Speech’ (1982) 130 University of Pennsylvania Law Review 591–645; Alexander Meicklejohn, Free Speech and Its Relation to Self-Government (Harper, 1948).

14 See Mark R Reiff, ‘Why Some People Think Fascism is the Greatest Expression of Democracy Ever Invented’ The Conversation (7 November 2022).

15 See Matteo Bonotti and Jonathan Seglow, ‘Freedom of Speech: A Relational Defence’ (2022) 48 Philosophy and Social Criticism 515–29.

16 See Frederick F Schauer, Free Speech: A Philosophical Inquiry (Cambridge University Press, 1982) 15–34.

17 See JS Mill, On Liberty (Cambridge University Press, 1989) 19–55.

18 See Abrams v. United States, 250 U.S. 616 (1919).

19 See Matteo Bonotti and Jonathan Seglow, Free Speech (Polity Press, 2021) 10–16, 12–13; David O Brink, ‘Millian Principles, Freedom of Expression, and Hate Speech’ (2001) 7 Legal Theory 119–57, 122–25.

20 See, e.g., Steven J Heyman, Free Speech and Human Dignity (Yale University Press, 2008); C Edwin Baker, Human Liberty and Freedom of Speech (Oxford University Press, 1989).

21 See, e.g., Steven H Shiffrin, Dissent, Injustice, and the Meanings of America (Princeton University Press, 1999); Fiss, Liberalism Divided (n 5).

22 See Bonotti and Seglow, ‘Freedom of Speech’ (n 15); Bonotti and Seglow, Free Speech (n 19) 8–30.

23 See generally Bonotti and Seglow, Free Speech (n 19); Martin H Redish, ‘The Value of Free Speech’ (1982) 130 University of Pennsylvania Law Review 591–645.

24 See Baker, Human Liberty and Freedom of Speech (n 20).

25 See Steven H Shiffrin, The First Amendment, Democracy, and Romance (Harvard University Press, 1990) 5.

26 See Seana Valentine Shiffrin, ‘A Thinker-Based Approach to Freedom of Speech’ (2011) 27(2) Constitutional Commentary 283–307.

27 See James Weinstein, Hate Speech, Pornography, and the Radical Attack on Free Speech Doctrine (Routledge, 1999) 7.

28 See Steven J Heyman, Free Speech and Human Dignity (Yale University Press, 2008).

29 See Thomas I Emerson, The System of Freedom of Expression (Vintage, 1970).

30 See Matthew H Kramer, Freedom of Expression as Self-Restraint (Oxford University Press, 2021) 152.

31 See, e.g., Schauer, Free Speech (n 16) 5–7ff.

32 Ibid.

33 See, e.g., Alvin I Goldman and Daniel Baker, ‘Free Speech, Fake News, and Democracy’ (2019) 18 First Amendment Law Review 66–141; Weinstein, ‘Participatory Democracy as the Central Value’ (n 13); James Weinstein, ‘Extreme Speech, Public Order, and Democracy’, in Ivan Hare and James Weinstein (eds), Extreme Speech and Democracy (Oxford University Press, 2009) 23–61; Lee C Bollinger, ‘Free Speech and Intellectual Values’ (1982-1982) 92 Yale Law Journal 438–73; Robert Bork, ‘Neutral Principles and Some First Amendments Problems’(1971) 47 Indiana law Journal 1–35, 20–35.

34 See Goldman and Baker, ‘Free Speech, Fake News, and Democracy’(n 33) 101–11, 125–39; Weinstein, ‘Extreme Speech, Public Order’ (n 33) 47–50; James Weinstein, ‘Hate Speech, Viewpoint Neutrality, and the American Concept of Democracy’ in Thomas R Hensley (ed.), The Boundaries of Freedom of Expression & Order in American Democracy (The Kent State University Press, 2001) 146–69, 147–48; Ronald Dworkin, ‘Is the Press Losing the First Amendment’ in A Matter of Principle (Harvard University Press, 1985) 381–97, 391.

35 For discussion of some of these see Bonotti and Seglow, Free Speech (n 19) 10–16; Schauer, Free Speech (n 16) 15–34.

36 See, e.g., C Edwin Baker, ‘Hate Speech’ in Michael Herz and Peter Molnar (eds), The Content and Context of Hate Speech: Rethinking Regulation and Responses (Cambridge University Press, 2012) 57–80; Erwin Chemerinsky and Howard Gillman, Free Speech on Campus (Yale University Press, 2017) 82–110.

37 See, e.g., Bianca Cepollaro, Maxime Lepoutre, and Robert Mark Simpson, ‘Counterspeech’ (2023) 18 Philosophy Compass 1–11.

38 See generally Mark R Reiff, Exploitation and Economic Justice in the Liberal Capitalist State (Oxford University Press, 2013); Mark R Reiff, ‘Two Theories of Economic Liberalism’ (2017) 10 The Adam Smith Review (Routledge) 189- 214.

39 See, e.g., William D Cohan, ‘How Loud Billionaires Convert Their Wealth into Power’ The New York Times (February 5, 2024); Catherine A MacKinnon, ‘Weaponizing the First Amendment: An Equality Reading’ (2020) 106 Virginia Law Review 1223–83, 1223–24.

40 This is a common point made by ‘free marketplace of ideas’ critics. See, e.g., Jerome A Barron ‘Access to the Press—A New First Amendment Right’ (1967) 80 Harvard Law Review 1641–78; Baker, Human Liberty and Freedom of Speech (n 20) 4–5.

41 See Steven Shiffrin, ‘Dissent, Democratic Participation, and First Amendment Methodology’ (2011) 97 Virginia Law Review 559–65, 562–63.

42 See Baker, Human Liberty and Freedom of Speech (n 20) 4.

43 See, e.g., Weinstein, Hate Speech, Pornography, and the Radical Attack (n 27) 5.

44 See, e.g., Renaud de la Brosse, ‘Political Propaganda and the Plan to Create a “State for All Serbs:” Consequences of Using the Media for Ultra-Nationalist Ends’ (2003) Report Compiled at the Request of the Office of the Prosecutor of the International Criminal Tribunal for the Former Yugoslavia.

45 See, e.g., Renée Loth, ‘The Trolls are Unleashed. What are Social Media Companies Doing About It?’ The Boston Globe (21 February 2018); Adam Goldman, ‘Justice Dept. Accuses Russians of Interfering in Midterm Elections’ The New York Times (19 October 2018); Evan Osnos, ‘Can Mark Zuckerberg Fix Facebook before It Breaks Democracy?’ The New Yorker (17 September 2018).

46 See Robert S Mueller, III, ‘Report on the Investigation into Russian Interference in the 2016 Presidential Election’ (March 2019) U.S. Department of Justice (Washington, DC); Ellen Nakashima, Shane Harris, Josh Dawsey, and Anne Gearan, ‘Senior Intelligence Official Told Lawmakers that Russia Wants to See Trump Reelected’ The Washington Post (21 February 2020); Julian E Barnes and Adam Goldman, ‘Russia Trying to Stoke U.S. Racial Tensions Before Election, Officials Say’ The New York Times (10 March 2020).

47 McKay Coppins, ‘The Billion-Dollar Disinformation Campaign to Reelect the President’ The Atlantic (10 February 2020). See also Evelyn N Farkas, ‘Russia is Interfering in Our Elections Again. And Trump Supporters are Emulating Russian Tactics’ The Washington Post (17 May 2020); Tiffany Hsu and Steven Lee Myers, ‘China’s Advancing Efforts to Influence the U.S. Election Raise Alarms’ The New York Times (1 April 2024); Joseph Menn, Aaron Schaffer, Naomi Nix, and Clara Ence Morse, ‘Chinese Propaganda Accounts Found by Meta Still Flourish on X’ The Washington Post (16 February 2024).

48 See John Rawls, ‘The Powers of Citizens and Their Representation’, in Political Liberalism (New York: Columbia University Press, exp. ed. 1993, 1996, 2005) 47–88, 63–64 & n 19.

49 See, e.g., John H Wigmore, Evidence (Little Brown, 1940) vol. 5, sec. 1467, 29.

50 See Goldman and Baker, ‘Free Speech, Fake News, and Democracy’ (n 33) 86–91.

51 See Baker, Human Liberty and Freedom of Speech (n 20) 6, 12–14.

52 See Ronald Dworkin, ‘What Rights Do We Have?’ in Taking Rights Seriously (Harvard University Press, 1977) 266–78.

53 See, e.g., Baker, Human Liberty and Freedom of Speech (n 20).

54 See Isaiah Berlin, ‘Two Concepts of Liberty’ in Henry Hardy (ed.) Liberty (Oxford University Press, 2002) 166–217.

55 See Rae Langton, ‘Lies and Back-Door Lies’ (2021) 130 Mind 251–58, 252.

56 See Berlin, ‘Two Concepts of Liberty’ (n 54).

57 See Reiff, ‘Neutrality and Excellence’ (n 8).

58 See Reiff, In the Name of Liberty (n 4) 14, 86.

59 See Fiss, Liberalism Divided (n 5) 4–5 (characterising the free speech debate as primarily driven by the perceived conflict between equality and liberty).

60 See, e.g., Reiff, In the Name of Liberty (n 4); Mark R Reiff ‘Left Libertarianism for the Twenty-First Century’ (2023) 2 Journal of Social and Political Philosophy 191–211.

61 See Paul Krugman, ‘Death and Tax Cuts’ The New York Times (24 February 2017).

62 See James Weinstein, ‘Hate Speech Bans, Democracy, and Political Legitimacy’ (2017) 32 Constitutional Commentary 527–84.

63 See, e.g., Alan Thomas, Republic of Equals: Predistribution and Property-Owning Democracy (Oxford University Press, 2017), xv (‘the purpose of this book is to construct and defend a liberal-republican theory of justice’); Philip Pettit, Republicanism: A Theory of Freedom and Government (Oxford University Press, 1997); Philip Pettit, On the People’s Terms: A Republican Theory and Model of Democracy (Cambridge University Press, 2012).

64 See, e.g., Robert S Taylor, Exit Left: Markets and Mobility in Republican Thought (Oxford University Press, 2017); Nicholas Vrousalis, Exploitation as Domination: What Makes Capitalism Unjust (Oxford University Press, 2023). See also further discussion of this point in Reiff, In the Name of Liberty (n 4) 88–94.

65 See, e.g., Suzanne Whitten, ‘Critical Republicanism and the Discursive Demands of Free Speech’ (2023) 49 Philosophy and Social Criticism 856–80; Suzanne Whitten, A Republican Theory of Free Speech (Palgrave Macmillan, 2022).

66 Note that I do not disagree that our current social practices and behaviour often privilege the speech of white men over that of women and people of colour. I merely disagree that this is something that is primarily the business of the right to free speech to remedy.

67 See Whitten, ‘Critical Republicanism and the Discursive Demands’ (n 65).

68 See Weinstein, Hate Speech, Pornography, and the Radical Attack (n 27).

69 See John Rawls, A Theory of Justice (Harvard University Press, 1971, rev. ed. 1999) 340, 347–50, 380–86.

70 See also Reiff, In the Name of Liberty (n 4) 58, 86–94, 230–38.

71 For more on what is and is not included in my definition of arbitrariness and why, see Reiff, In the Name of Liberty (n 4) 89–93; Reiff, ‘Left Libertarianism for the Twenty-First Century’ (n 60) 201–02.

72 See Ava Sasani, ‘“A Concern for Everyone”: Tennessee Poised to Ban Pride Flags in Schools’ The Guardian (3 March 2024).

73 See Nathaniel Frank, ‘What the Science Says About “Don’t Say Gay” and Young People’ The New York Times (20 April 2023).

74 For example, the US Administrative Procedure Act, section 10(e), provides that the courts ‘shall … set aside agency action … found to be . . . arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law’, 5 U.S.C. 1009 (1946). And since the enactment of that statute in 1946, countless administrative and executive decisions have been reviewed under that standard alone. See also Garfield v. United States ex rel. Goldsby, 211 U.S. 249, 262 (1908) (‘there is no place in our constitutional system for the exercise of arbitrary power’); Yick Wo v. Hopkins, 118 U.S. 356, 370 (1886) (Our institutions ‘do not mean to leave room for the play and action of purely personal and arbitrary power’).

75 See Mark R Reiff, ‘Twenty-One Statements about Political Philosophy: An Introduction and Commentary on the State of the Profession’ (2018) 41 Teaching Philosophy 65–115, 80.

76 See generally Andrew T Kenyon, ‘Assuming Free Speech’ (2014) 77 The Modern Law Review 379–408.

77 See Reiff, In the Name of Liberty (n 4).

78 See Robert Nozick, Anarchy, State, and Utopia (Basic Books, 1974), 87, 110–13; Reiff, In the Name of Liberty (n 4) 46–49; Reiff, ‘Left Libertarianism for the Twenty-First Century’ (n 71).

79 See Neal Kumar Katyal, ‘Conspiracy Theory’ (2002-2203) 112 Yale Law Journal 1307–98.

80 See Protocol Additional to the Geneva Conventions of 12 August 1949, and Relating to the Protection of Victims of International Armed Conflicts (Protocol I), June 8, 1977, art. 51(5)(b).

81 See the discussion of ‘critical risk’ in Mark R Reiff, Punishment, Compensation, and Law: A Theory of Enforceability (Cambridge University Press, 2005) 92–95, 109–10.

82 See Reiff, In the Name of Liberty (n 4).

83 See TM Scanlon, ‘A Theory of Freedom of Expression’ in The Difficulty of Tolerance: Essays in Political Philosophy (Cambridge University Press, 2003) 6–25.

84 See Berlin, ‘Two Concepts of Liberty’ (n 54).

85 See Post, ‘Participatory Democracy and Free Speech’ (n 13) 480.

86 See Leslie Kendrick, ‘Free Speech as a Special Right’ (2017) 45 Philosophy and Public Affairs 87–117, 113–66.

87 See C Edwin Baker, ‘Autonomy and Free Speech’ (2011) 27 Constitutional Commentary 251–82.

88 See Baker, Human Liberty and Freedom of Speech (n 20) 3.

89 Hillel Steiner, An Essay on Rights (Blackwell, 1994) 52–54.

90 See Mark R Reiff, On Unemployment, Volume I: A Macro Theory of Economic Justice (Palgrave Macmillan, 2015) 125–27.

91 See Reiff, In the Name of Liberty (n 4) 86–87; Charles Taylor, ‘What’s Wrong with Negative Liberty?’ in Philosophy and the Human Sciences: Philosophical Papers 2 (Cambridge University Press, 1985) 211–29.

92 See Bonotti and Seglow, Free Speech (n 19) 16–18.

93 See Reiff, Exploitation and Economic Justice (n 38) 84–93; Nozick (n 78) 262–65; Charles Fried, Contract as Promise (Harvard University Press, 1981) 96–99.

94 See Mark R Reiff, ‘Proportionality, Winner-Take-All, and Distributive Justice’ (2009) 8 Politics, Philosophy, and Economics 5–42.

95 For discussion of the some of the problems here and the attempts to get around them, see Alexander Brown and Adriana Sinclair, Hate Speech Frontiers (Cambridge University Press, 2024); Alexander Brown, ‘What is Hate Speech? Part I: The Myth of Hate’ (2017) 36 Law and Philosophy 419–68.

96 See, e.g., Katherine Gelber, ‘Differentiating Hate Speech: A Systemic Discrimination Approach’ (2021) 24 Critical Review of International Social and Political Philosophy 393–414.

97 See Rawls, A Theory of Justice (n 69) 266.

98 See generally Mateo Bonotti, ‘Religion, Hate Speech and Non-Domination’ (2017) 17 Ethnicities 259–74, 272–73.

99 See Mark R Reiff, ‘A Philosopher Argues Why No One Has the Right to Refuse Services to LGBT People’ The Conversation (25 July 2017).

100 See Reiff, ‘Neutrality and Excellence’ (n 8) 280–81. As best I can determine, this distinction is also what Philip Pettit is intending to refer to when he distinguishes ‘unhindered’ speech from ‘protected’ speech. See Philip Pettit, ‘Two Concepts of Free Speech’ in Jennifer Lackey (ed.), Academic Freedom (Oxford University Press, 2018) 61–81. For an extended discussion of the Hohfeld’s long-standing and almost universally accepted system of categorisation of juridical relations, which Pettit’s distinction echoes but for some reason never mentions, see Mathew H Kramer, A Debate Over Rights (Oxford University Press, 1998) 7–60. See also Kramer, Freedom of Expression as Self-Restraint (n 30) 10–16.

101 For a much fuller presentation of this point, see Reiff, ‘Neutrality and Excellence’ (n 8) 280–81.

102 See, e.g., Seana Valentine Shiffrin, Speech Matters: On Lying, Morality, and the Law (Princeton University Press, 2014) 5–46. Shiffrin, however, protects certain lies despite their being outside the scope of the right to free speech, see Langton (n 55), while my approach shows why these cases are not exceptions to the right but instantiations of it.

103 See Stanley Fish, The First: How to Think About Hate Speech, Campus Speech, Religious Speech, Fake News, Post-Truth, and Donald Trump (Simon & Schuster, 2019) 31–61; Jeffrey W Howard, ‘Free Speech and Hate Speech’ (2019) 22 Annual Review of Political Science 93–109, 95; Alexander Brown, Hate Speech Law: A Philosophical Examination (Routledge, 2015) ch. 1.

104 See generally Jeremy Waldron, The Harm in Hate Speech (Harvard University Press, 2012).

105 For a recent discussion of this point and reference to other examples, see Enes Kulenović, ‘Should Democracies Ban Hate Speech? Hate Speech Laws and Counterspeech’ (2022) Ethical Theory and Moral Practice sec. 3; Reverend William Barber, ‘The Racist Murders in Jacksonville Didn’t Happen in a Vacuum. Words Came First’ The Guardian (29 August 2023); Tasnim Nazeer, ‘The Anti-Muslim Rhetoric of Rightwing Politicians is Fuelling Hate Crime – I’ve Experienced It Myself’ The Guardian (26 February 2024).

106 See, e.g., Nathan P Kalmoe, ‘Fueling the Fire: Violent Metaphors, Trait Aggression, and Support for Political Violence’ (2014) 31 Political Communication 545–63; Karsten Müller and Carlo Schwarz, ‘From Hashtag to Hate Crime: Twitter and Anti-Minority Sentiment’ SSRN (http://dx.doi.org/10.2139/ssrn.3149103) (24 July 2020).

107 See Constance Grady, ‘The Alice Walker Anti-Semitism Controversy, Explained’ Vox (20 December 2018); Alexandra Alter, ‘Alice Walker, Answering Backlash, Praises Anti-Semitic Author as “Brave”’ The New York Times (21 December 2018).

108 See, e.g., Julia Angwin, ‘Few are Addressing One of Social Media’s Greatest Perils’ The New York Times (6 May 2023).

109 See Carl Schmitt, The Concept of the Political, trans. George Schwab (University of Chicago Press, 1996).

110 See Reiff, ‘Neutrality and Excellence’ (n 8).

111 See, e.g., Leslie Kendrick, ‘Lies and Free Speech Values’ (2018) 38 Law and Philosophy 495–506 (discussing Shiffrin, Speech Matters).

112 See Mari J Matsuda, ‘Public Response to Racist Speech: Considering the Victim's Story’ (1989) 87 Michigan Law Review 2320–81.

113 See Rawls, A Theory of Justice (n 69) 190–94.

114 See Reiff, ‘Neutrality and Excellence’ (n 8).

115 See Reiff, ‘Neutrality and Excellence’ (n 8) 277–78.

116 See, e.g., Schauer, Free Speech (n 16) 80–86.

117 See, e.g., David Smith, ‘'Words of a Dictator': Trump's Threat to Deploy Military Raises Spectre of Fascism’ The Guardian (2 June 2020); Richard Wolffe, ‘Trump Has Reached the “Mad Emperor” Stage, and It's Terrifying to Behold’ The Guardian (2 June 2020).

118 See, e.g., Aaron Blake, ‘The Trump White House’s Dangerously Authoritarian Response to Criticism’ The Washington Post (10 October 2017); Peter Baker and Cecilia Kang, ‘Trump Threatens NBC over Nuclear Weapons Report’ The New York Times (11 October 2017); Greg Sargent, ‘In New Interview, Trump Openly Rages at Checks on His Authoritarianism’ The Washington Post (3 November 2017); Rebecca Ratcliffe, ‘Civil Rights “Under Serious Attack” Across the Globe’ The Guardian (27 November 2018).

119 See, e.g., Shaun Walker, ‘Hungarian Journalists Fear Coronavirus Law May Be Used to Jail Them’ The New York Times (3 April 2020).

120 See Stewart L Udall, ‘Robert Frost’s Last Adventure’ The New York Times (11 June 1972).

121 See 47 U.S.C. § 230.

122 See John Rawls, Political Liberalism (Columbia University Press, 1993, 1996) 54ff.

123 See Fareed Zakaria, ‘Liberals Should Tread Carefully When Confronting Trumpism’ The Washington Post (29 March 2024).

124 See, e.g., Leroy Edozien, ‘Determining the Appropriate Standard of Care: Even “Accepted Medical Practice” May Be Negligent’ (2001) 7 Clinical Risk 103–06; Kenneth S Abraham, ‘Custom, Noncustomary Practice, and Negligence’ (2009) 109 Colum L Rev 1784–822, 1804–05.

125 See Craig Silverman, ‘How the Bizarre Conspiracy Theory Behind “Pizzagate” Was Spread’ BuzzFeed (4 November 2016); Eli Rosenberg, ‘Alex Jones Apologizes for Promoting “Pizzagate” Hoax’ The New York Times (25 March 2017).

126 See Aria Bendix, ‘“Pizzagate” Shooter to Serve Four Years in Jail’ The Atlantic (22 June 2017). See also Adrian Horton, ‘After Truth; How Ordinary People are ‘Radicalized’ by Fake News’ The Guardian (19 March 2020).

127 See Vanessa Romo, ‘Sandy Hook Victim's Father Wins Defamation Suit; Alex Jones Sanctioned’ NPR (18 June 2019); Michael Gold and Tyler Pager, ‘Sandy Hook Victim’s Father Dies in Apparent Suicide in Newtown’ The New York Times (25 March 2019); Tim Adams, ‘Sandy Hook’s Tragic Legacy: Seven Years On, a Loving Father is the Latest Victim’ The Guardian (31 March 2019). Note also that despite the entry of a mammoth judgment for damages against Jones, the plaintiffs have yet to receive a cent, indicating how empty the promise of ex post facto remedies can be in both deterring this kind of behaviour and compensating victims when it nevertheless occurs. See Jada Yuan, ‘“The Truth vs. Alex Jones”: How Sandy Hook Lies Got Peddled for Profit’ The Washington Post (26 March 2024).

128 See, e.g., Peter Stone, ‘Texas Fracking Billionaire Brothers Fuel Rightwing Media with Millions of Dollars’ The Guardian (5 September 2023).

129 See Graham Readfearn, ‘Scientific Journal Retracts Article that Claimed No Evidence of Climate Crisis’ The Guardian (25 August 2023).

130 See, e.g., Coral Davenport and Eric Lipton, ‘How G.O.P. Leaders Came to View Climate Change as Fake Science’ The New York Times (3 June 2017); Aliya Uteuova, ‘Nearly 15% of Americans Don’t Believe Climate Change is Real, Study Finds’ The Guardian (14 February 2024).

131 See, e.g., Brad Plumer and Coral Davenport, ‘Science Under Attack: How Trump is Sidelining Researchers and Their Work’ The New York Times (28 December 2019); Steven Salzberg, ‘How Anti-Vax Activists Use Conspiracy Theories to Spread Fear of Vaccines’ Forbes (3 February 2020).

132 See, e.g., Elizabeth Dwoskin, ‘A Quarter of Trump’s 6,081 Facebook Posts Last Year Featured Misinformation or Extreme Rhetoric’ The Washington Post (18 February 2021).

133 See, e.g., Rosalind S Helderman and Elise Viebeck, ‘“The Last Wall”: How Dozens of Judges across the Political Spectrum Rejected Trump’s Efforts to Overturn the Election’ The Washington Post (12 December 2020); Trump v. Boockvar, Civil Action No. 4:20-CV-02078, Memorandum Opinion of US District Judge Matthew W Brann (Middle Dist. Pa) (21 November 2020); Aaron Blake, ‘The Most Remarkable Rebukes of Trump’s Legal Case: From the Judges He Hand-Picked’ The Washington Post (14 December 2020); Katie Benner and Michael S Schmidt, ‘Barr Acknowledges Justice Dept. Has Found No Widespread Voter Fraud’ The New York Times (1 December 2020); Ann Gerhart, ‘Election Results Under Attack: Here are the Facts’ The Washington Post (14 January 2021); Amy Davidson Sorkin, ‘The Supreme Court Rejects Texas’s Shameful Lawsuit, But There Has to be a Reckoning’ The New Yorker (12 December 2020).

134 See, e.g., David E Sanger, Matt Stevens, and Nicole Perlroth, ‘Election Officials Directly Contradict Trump on Voting System Fraud’ The New York Times (12 November 2020).

135 See Richard Posner, Not a Suicide Pact (Oxford University Press, 2006).

136 See Lena H Sun, Lauren Weber, and Hayden Godfrey, ‘Doctors Who Put Lives at Risk with Covid Misinformation Rarely Punished’ The Washington Post (26 July 2023).

137 See Naja Bentzen, ‘Trump's Disinformation “Magaphone”: Consequences, First Lessons and Outlook’ European Parliament Research Service (February 2021) (https://www.europarl.europa.eu/RegData/etudes/BRIE/2021/679076/EPRS_BRI(2021)679076_EN.pdf).

138 See Kevin Roose, ‘Reddit Limits Noxious Content by Giving Trolls Fewer Places to Gather’ The New York Times (25 September 2017); Eshwar Chandrasekharan, et al., ‘You Can’t Stay Here: The Efficacy of Reddit’s 2015 Ban Examined Through Hate Speech’ (November 2017) 1(2) ACM Transactions on Computer-Human Interaction Article 31.

139 See Kari Paul, ‘A Few Rightwing “Super-Spreaders” Fueled Bulk of Election Falsehoods, Study Says’ The Guardian (5 March 2021).

140 See Glenn Kessler, ‘Trump Made 30,573 False or Misleading Claims as President. Nearly Half Came in His Final Year’ The Washington Post (23 January 2021); Andrew S Ross and Damian J Rivers, ‘Discursive Deflection: Accusation of “Fake News” and the Spread of Mis- and Disinformation in the Tweets of President Trump’ (2018) Social Media + Society 1–12.

141 See Hannah Denham, ‘These are the Platforms that Have Banned Trump and His Allies’ The Washington Post (13 January 2021); Tony Romm and Elizabeth Dwoskin, ‘Twitter Purged More Than 70,000 Accounts Affiliated with QAnon Following Capitol Riot’ The Washington Post (11 January 2021); Jack Nicas and Davey Alba, ‘Amazon, Apple and Google Cut Off Parler, an App That Drew Trump Supporters’ The New York Times (9 January 2021); Guardian Staff and Agency, ‘Twitter Suspends Marjorie Taylor Greene, QAnon-Backing Republican’ The Guardian (18 January 2021).

142 See Elizabeth Dwoskin and Craig Timberg, ‘Misinformation Dropped Dramatically the Week after Twitter Banned Trump and Some Allies’ The Washington Post (16 January 2021).

143 See Arwa Mahdawi, ‘How Many Anti-Vaxxers Does It Take to Misinform the World? Just Twelve’ The Guardian (30 March 2021).

144 Deplatforming had previously proved effective with regard to limiting Isis and Al Qaeda recruiting and the radicalisation of sympathetic individuals by other jihadists, but due to a lack of other examples, it was thought that this tool would be less effective when applied to more home-grown forms of extremism. See Matt Bradley and Mo Abbas, ‘Is Deplatforming Extremists Effective or Dangerous? Experts Weigh In’ NBC News (18 January 2021); Shawn Boburg, ‘From Corporate America to Conspiracy Theory Promotion: How a Minnesota Man Made a Career Out of Anonymously Amplifying Dark Plots’ The Washington Post (7 July 2021).

145 See Colin Dickey, ‘From Sound of Freedom to Ron DeSantis: How QAnon’s Crazy Conspiracy Theories Went Mainstream’ The Guardian (16 August 2023).

146 See Stuart A Thompson, ‘Musk Lifted Bans for Thousands on Twitter. Here’s What They’re Tweeting’ The New York Times (22 December 2022). See also Josh Taylor, ‘Bots on X Worse Than Ever According to Analysis of 1m Tweets During First Republican Primary Debate’ The Guardian (8 September 2023).

147 See, e.g., Julie Carrie Wong, ‘Tech Giants Struggle to Stem “Infodemic” of False Coronavirus Claims’ The Guardian (10 April 2020).

148 See Naomi Nix and Sarah Ellison, ‘Following Elon Musk’s Lead, Big Tech is Surrendering to Disinformation’ The Washington Post (1 September 2023).

149 See Jan Wolfe, ‘“He Invited Us”: Accused Capitol Rioters Blame Trump in Novel Legal Defense’ Reuters (2 February 2021); Olivia Rubin, Alexander Mallin, and Alex Hosenball, ‘“Because President Trump Said To”: Over a Dozen Capitol Rioters Say They were Following Trump's Guidance’ ABC News (9 February 2021) (some of them are accused of perpetrating the attack's most violent crimes); Dan Barry, Mike McIntire, and Matthew Rosenberg, “‘Our President Wants Us Here”: The Mob That Stormed the Capitol’ The New York Times (10 November 2021).

150 See David O Brink, ‘Millian Principles, Freedom of Expression, and Hate Speech’ (n 19) 119.

151 For a similar view, albeit based on different grounds, see Étienne Brown, ‘Free Speech and the Legal Prohibition of Fake News’ (2023) 49 Social Theory and Practice 29–55.

152 See Tierney Sneed, ‘Inside the Effort to Disbar Attorneys Who Backed Bogus Election Lawsuits’ CNN (10 March 2022).

153 See Mark P Gergen, ‘Negligent Misrepresentation as Contract’ (2013) 101 California Law Review 953–1011.

154 See Kramer, Freedom of Expression as Self-Restraint (n 30) 101–07; Joseph M Perillo, ‘The Origins of the Objective Theory of Contract Formation and Interpretation’ (2001–2001) 69 Fordham Law Review 427–77, 431.

155 See generally Dan B Dobbs, Paul T Hayden, and Ellen M Bublick, Hornbook on Torts (West Publishing, 2nd edn, 2016), ch. 37.

156 See Reiff, ‘Two Theories of Economic Liberalism’ (n 38).

157 See Shiffrin, Speech Matters (n 102) ch. 4.

158 See, e.g., Bork, ‘Neutral Principles and Some First Amendments Problems’ (n 33).

159 See Naomi Nix, Michael Scherer, and Jeremy B Merrill, ‘As Meta Flees Politics, Campaigns Rely on New Tricks to Reach Voters’ The Washington Post (21 April 2024); Cecillia Kang and Mike Issac, ‘Defiant Zuckerberg Says Facebook Won’t Police Political Speech’ The New York Times (17 October 2019).

160 Of course, some theorists have argued that for this very reason, enhancements for hate crimes are an unacceptable infringement of free expression. But the fact that such enhancements have been repeatedly upheld indicate that we do generally recognise that toleration and neutrality are not unlimited even in a liberal society, and that not only procedural limits but also substantive limits apply. See generally Frederick M Lawrence, ‘Resolving the Hate Crimes/Hate Speech Paradox: Punishing Bias Crimes and Protecting Racist Speech’ (1992) 68 Notre Dame Law Review 673–722.

161 See John Bruce Lewis and Bruce L Ottley, ‘New York Times v. Sullivan at 50: Despite Criticism, the Actual Malice Standard Still Provides Breathing Space for Communications in the Public Interest’ (2014) 64 DePaul Law Review 1–64.

162 See Bonotti and Seglow, ‘Freedom of Speech: A Relational Defence’ (n 15) 516. Even John Rawls buys into this distinction. See Samuel Freeman, Rawls (Routledge, 2007) 210–11.

163 See, e.g., Ramesh Ponnuru, ‘The Big Winner of the 2024 GOP Primaries Was the “Big Lie”’ The Washington Post (24 January 2024); Jim Rutenberg and Steven Lee Myers, ‘How Trump’s Allies Are Winning the War Over Disinformation’ The New York Times (17 March 2024); Luke Broadwater and Alan Feuer, ‘Jan. 6 Panel Tracks How Trump Created and Spread Election Lies’ The New York Times (13 June 2022).

164 See, e.g., Andrew Higgins, ‘The Art of the Lie? The Bigger the Better’ (10 June 2021).

165 See, e.g., Lauren Weber and Sabrina Malhi, ‘Women are Getting Off Birth Control Amid Misinformation Explosion’ The Washington Post (21 March 2024).

166 See Schenck v. United States, 249 U.S. 47, 52 (1919).

167 See Stanley Fish, ‘Boutique Multiculturalism, or Why Liberals Are Incapable of Thinking about Hate Speech’ (1997) 23 Critical Inquiry 378–95, 392. Note that while Fish claims this is a lesson liberals will never learn, he is using the terms ‘liberal’ and ‘liberalism’ in the more colloquial sense, to refer to left-leaning theorists and their followers, not as I am doing here, to refer to a family of political theories that accept certain fundamental presuppositions, including the presuppositions of toleration and neutrality, but cash them out in various ways, ways that can fall to the right as well as the left.

168 See, e.g., Taylor Lorenz, ‘Meta is Failing to Curb Anti-Trans Hate, New Report Says’ The Washington Post (27 March 2024); Weronika Strzyżyńska, ‘Meta and Google Accused of Restricting Reproductive Health Information’ The Guardian (27 March 2024); Jim Rutenberg and Kate Conger, ‘Elon Musk is Spreading Election Misinformation, but X’s Fact Checkers are Long Gone’ The New York Times (29 January 2024); Kari Paul, ‘Reversal of Content Policies at Alphabet, Meta and X Threaten Democracy, Warn Experts’ The Guardian (7 December 2023); Jason Wilson, ‘Rightwing Personalities Use X to Bring Antisemitic Theories to Light in US’ The Guardian (21 November 2023); Chris Michael, ‘Meta Allows Facebook and Instagram Ads Saying 2020 Election Was Rigged’ The Guardian (15 November 2023).

169 See, e.g., Naomi Nix, ‘Hundreds of Groups Urge Big Tech CEOs to Step Up Fight against AI-Fueled Lies’ The Washington Post (9 April 2024); Rachel Leingang, ‘“Disinformation on Steroids”: Is the US Prepared for AI’s Influence on the Election?’ The Guardian (26 February 2024); Naomi Nix, ‘Oversight Board Rebukes Meta’s Policies after Altered Biden Video Spreads: The Board Called the Policies on Deepfakes ‘Incoherent’ and ‘Confusing’, Amid AI Fears’ The Washington Post (5 February 2024); Cat Zakrzewski and Caroline O'Donovan, ‘Lawmakers Demand Answers from Bezos about Election Misinformation on Alexa’ The Washington Post (19 October 2023) (when asked whether the 2020 election was stolen, Alexa answered that there was ‘a massive amount of election fraud’).

170 See Cathy Park Hong ‘How Fake News is Hatching in Immigrant Communities’ The New York Times (20 December 2020).

171 See also Paul Krugman, ‘Facts Have a Well-Known Liberal Bias’ The New York Times (8 December 2017).