Statue of Jon Stewart Mill by Thomas Woolner, London. Flickr

Problems with Mill’s Masterpiece

Bo Winegard
Bo Winegard
18 min read

The condition of human life is such that we must of necessity be restrained and compelled by circumstances in nearly every action of our lives. Why, then, is liberty, defined as Mr. Mill defines it, to be regarded as so precious?
~James Fitzjames Stephen

Of all the works written in defense of free speech, and of liberalism more broadly, few have been as celebrated as John Stuart Mill’s On Liberty, which is widely regarded as the definitive source for free speech advocacy. Mill advances the audacious argument that “the sole end for which mankind are warranted … in interfering with the liberty of action of any of their number, is self-protection.” Thus, he demands “liberty of … conscience … in the most comprehensive sense,” of which freedom of expression and the press are “practically inseparable” parts. Mill champions an expansive, almost absolute position on freedom of speech, specifically considering and then rejecting appeals to moderation with the claim that “unless the reasons [for free expression] are good for an extreme case, they are not good for any case.”

Furthermore, Mill not only attacks suppression and censorship backed by government force, but also suppression and censorship promoted by pressures for conformity among the public, lamenting that such pressures are often as effective as prison: “In respect to all persons but those whose pecuniary circumstances make them independent of the good will of other people, opinion, on this subject, is as efficacious as law; men might as well be imprisoned, as excluded from the means of earning their bread.”

Not without reason, then, many modern proponents of free speech venerate Mill’s work. However, like many classics, it is almost certainly more often referenced than read. As a result, the excellence of its arguments has been exaggerated (outside of the scholarly literature, where many critical essays and books have been published), while its flaws have been ignored. This is a shame because free speech is a contentious topic that requires careful thought and consideration, not appeals to authority and reflexive deference. Although it offers an eloquent defense of liberty, Mill’s book is also full of vague and contradictory arguments, dubious claims, and tendentious declarations. On balance, its defense of free speech should be considered a failure since most of its chief assertions are either plausibly incorrect or apply to only limited domains of discourse.

A critical re-examination of Mill’s major claims ought to be useful, regardless of one’s own views about free speech. But it is particularly important at a time when personal liberty is under attack from illiberal and populist ideologies. And, as Mill himself would surely have argued, the first step to building a vigorous defense is diligent criticism, so that fragile arguments can be replaced by more resilient reasoning.


Mill was famously an advocate of the consequentialist ethical philosophy called utilitarianism. Succinctly, utilitarianism is the view that the most moral action or rule is that which produces the greatest happiness (utility) for the greatest number of people. Like any attempt to summarize a complicated philosophy in a single sentence, this is wildly simplistic, but it will have to suffice. Although Mill often couches his claims about liberty in absolutist language, he asserts that his framework ultimately rests upon utilitarian assumptions: “I regard utility as the ultimate appeal on all ethical questions.” This is important because it allows a skeptic to critique Mill on his own terms. According to his own principles, Mill succeeds or fails only insofar as his arguments are consistent with utilitarianism.

While Mill’s abstract principles are putatively utilitarian, his basic premise in On Liberty is more specific: “…the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others. His own good, either physical or moral, is not sufficient warrant.” Although this claim (known as the “harm principle”) is often quoted approvingly, it immediately raises a host of problems.

The first and most obvious of these is that most people approve of society exercising power over children, teenagers, and many others “for their own good.” In anticipation of this objection, Mill argues that children and “barbarians” do not belong to the “civilized community,” which consists only of rational adults (in certain civilizations), and are therefore exempt from the harm principle. In fact, Mill notes that “despotism” might be a “legitimate form of government” for those who are not part of this civilized community, so long as it is a despotism that works for their own good. One would certainly express these claims differently today (barbarian is rarely used unironically), but Mill’s distinction between rational adults and others is surely reasonable and defensible.

A second and more serious objection is that it is unclear what counts as “harm to others.” Since Mill is a self-professed utilitarian, his notion of harm should, I think, refer to any stimulus, event, action, et cetera which decreases another person’s utility and could therefore include anything from a punch and wrongful imprisonment to unpleasant odors and offensive words. If harm is understood this broadly, though, the harm principle becomes virtually worthless, since it would comprise many behaviors, and power could be “rightfully exercised over” people not only to prevent them from physically assaulting others, but also to prevent them from writing vulgar or disagreeable pamphlets or refusing to shower.

Of course, this is not the view of harm that Mill has in mind since he advocates for virtually unlimited speech. Therefore, a more charitable reading is that by harm Mill means immediate physical injury. But Mill does not justify this circumscribed notion of harm, which in any case seems to violate his own ethical principles—utilitarianism does not rest upon a metaphysical distinction between physical and psychological harm. (From strictly utilitarian reasoning, it is difficult to justify the distinction, since in many cases people would prefer physical to psychological suffering. A person may, for instance, prefer to be punched in the stomach than to live in a community in which their reputation had been destroyed by slander.)

A more precise statement of the legitimate use of power from a utilitarian perspective might be: “Power may be exercised over the individual only when that exercise, on average, increases the aggregate utility of the population.” Wording it this way reveals one of the unappealing features of utilitarianism, which critics have assailed since its origin: It seems inconsistent with the idea of inviolable rights and principles (save for the utility principle), since such rights and principles are not based on the careful calculation of consequences.

Mill argues that this is a superficial inconsistency, and that utilitarianism, properly understood, is congruent with rights and principles. He might respond that his harm principle better maximizes utility in the long run than would more impulsive and shortsighted approaches that condone exercising power to deter the expression of obnoxious opinions or dissent. But he does not make that case in On Liberty, so we are left with an incomplete argument. It is difficult to know how Mill got to the harm principle and even more difficult to see how it can be defended from utilitarian principles.

Mill offers four arguments in support of an expansive view of free speech, which can be reduced to three. I will address them in turn.

The silencing of discussion is an assumption of infallibility

Mill makes a moral claim that silencing speech is always wrong and an empirical claim that silencing is always based on the assumption of infallibility. These arguments are probably the most powerful, at least rhetorically, and appeal to a spirit of humility. Censorship and suppression are always based on at least an implicit belief in the correctness of the orthodoxy. Although this is rhetorically powerful, it is false and often irrelevant for four reasons:

(1) Near certainty about specific beliefs does not in fact require an assumption of infallibility, either narrowly or broadly conceived. In schools, children are taught various historical facts. About some of these facts, we are as close to certain as it is possible to be: “Germany participated in the First World War”; “John Fitzgerald Kennedy was assassinated”; “The Gulf of Tonkin Resolution passed in 1964.” Schools systematically exclude opposing views about these events and others and would almost certainly fire teachers who insisted on teaching or debating alternative ideas. This is not based on the elaborate premise that the orthodoxy is, in fact, infallible, but on the mundane premise that relative certainty about the world is possible in some domains and it would be wasteful and confusing to teach alternative views that are poorly supported.

It is also possible to make this distinction about certain moral claims. “Children should not be sold as slaves” is a claim about morality. Nevertheless, most people would probably agree to this assertion as enthusiastically as they would agree to an assertion about geometry or logic. One can have a long, entertaining, and probably fruitless quarrel about moral realism, but whatever one’s view about the objectivity of moral claims, the assertion that children should not be sold as slaves is clearly true in all ways that matter to a community. Agreement with this assertion, even with 100 percent certainty and the fervor that such certainty entails, does not suggest a belief in moral infallibility.

Again, it is possible to be certain enough about some moral and metaphysical claims to make alternative views taboo. Modern communities do, in fact, silence pro-slavery arguments. A teacher who defended slavery would be fired; people who tweeted in its favor would be booted from Twitter; essayists who promoted slavery would be shunned, mocked, and marginalized. These are not legal sanctions, but one of Mill’s primary concerns is the power of public opinion to suppress and silence. This is a clear case of silencing that most people, even people with expansive views of speech, would support based on certainty about a moral proposition which does not presuppose a belief in the infallibility of the orthodoxy; instead, it presupposes near certainty about a specific claim.

(2) It is often easier to be certain that particular theories or claims are false than that they are true. In many cases, certainty is impossible. It does not seem likely, for example, that anyone will ever know for certain what the optimal tax rate is, why consciousness exists, or how life began. It would be foolish to claim certainty for specific assertions about these (and many other) topics, and Mill is right to counsel the orthodoxy against smugly dismissing alternative views. Nevertheless, although certainty about specific claims and assertions is frequently illusory, and indicative of at least a kind of reflexive belief in the infallibility of experts, certainty about the falseness of various claims and assertions is often justified. For example, economists are justifiably certain that 100 percent is not the optimal tax rate; philosophers are justifiably certain that consciousness is not a gift from Zeus after a fiery sacrifice by a maenad in 355 BCE; and scientists are justifiably certain that life did not begin after a metaphysical hammer shot sparks down to Earth while fixing the stars in the sky.

(3) Communities and individuals are often required to make difficult decisions based on limited knowledge. Even if one were to grant Mill’s argument that suppression is always based on the illegitimate presumption of infallibility, it does not follow that suppression is immoral, illogical, and unwise. Before considering speech specifically, let us consider a more consequential restriction of personal liberty: Sentencing somebody to life in prison. One might argue, in the style of Mill, that sentencing people to prison requires the assumption of infallibility. After all, what could possibly justify using state power to remove somebody from society and force them to dwell for the rest of their life in an undesirable and wretched place other than absolute certainty? But of course, in the criminal justice system, absolute certainty, although a laudable goal, is generally impossible to attain.

The reason that society allows the criminal justice system to imprison people for life despite a degree of uncertainty is that it must weigh the costs and the benefits of its process. A criminal justice system that made no false positives because it insisted on infallibility would produce an endless stream of false negatives. Not only would the guilty walk the streets, depriving victims of justice, but crime rates would also be high because deterrence would be obsolete.

The same logic must hold for speech, especially if one accepts, as Mill does, a consequentialist framework that insists that outcomes are ultimately what determine the wisdom of a law or norm. If outcomes are what matter, surely one must weigh both the costs and the benefits of speech. And the suppression of speech, therefore, may be based not on a presumption of infallibility, but rather on a cost-benefit analysis, just as the criminal justice system is. Of course, this does not mean that suppression is, in fact, generally a good idea (I don’t think it is), but it does mean that Mill’s argument, although rhetorically powerful, is erroneous.

(4) Communities may silence information which they know is true because they think it is potentially dangerous or offensive. It is not clear that it is always better to propound ideas, thoughts, or hypotheses that are true—in some cases, communities defensibly suppress or discourage speech which is not only possibly but also very likely to be accurate. For an example of simple social decorum, consider open conversation about profoundly unattractive people. One can simultaneously recognize the truth of the general claim that it is incredibly difficult and unpleasant to look at some people while also recognizing that it is probably best to refrain from openly discussing the ugliness of specific people. Suppose that a professor pointed at a student and said, “And, of course, John Smith here is so ugly that it is painful to my eyes.” It is doubtful that people would respond sympathetically to him if he attempted to defend his assertion by noting that it is, in fact, true.

There are many cases of more consequence. Secrets about other people. Secrets about pathogens. Secrets about destructive weapons. Many communities consider it indecent to divulge private information about another person, even if it is true. The objection is ultimately that the information is more costly to the individual than it is beneficial to the community. Similarly, many governments and agencies keep secrets about bioterrorism, and plenty of accurate information about building weapons is hidden from the public. Suppressing these forms of speech is not, in fact, based on an assumption of infallibility. Mill’s assertion that “All silencing of discussion is an assumption of infallibility” is demonstrably false.

Silenced speech, although wrong, might contain some truth

Mill’s second major argument against silencing speech is more limited and judicious than the first and, therefore, more plausible and persuasive (though less striking). However, it is also vulnerable to many of the same objections. Although the orthodoxy is often wrong or only partially correct and so benefits from criticism, this does not mean that all criticism is healthy or valuable. Perhaps, for example, age-of-sexual-consent laws are too high or too low in some states, and the orthodoxy in those states would thus benefit from criticism and debate. It obviously does not follow that communities should be open to child-adult sex advocacy. This basic argument holds, I think, for almost every conceivable belief or assertion, although the force of public pressure applied against dissent will depend on its perceived harm and plausibility.

For example, the orthodox belief is that men and women are incarcerated at different rates largely because men commit crimes, especially violent crimes, at higher rates than women. This should be challenged. And Mill is certainly correct that some heterodox claims, even if largely incorrect, might contain some truth worth hearing. One might argue that, in fact, men do not commit crimes at higher rates than women, but that the criminal justice system is prejudiced against men. Although the first part of this argument is certainly false, the second part is plausible; and it may, in its clash with the orthodoxy, lead to a more complete picture of disparities in the criminal justice system.

Nevertheless, this hardly means that all contrary claims are edifying. Suppose, for example, a scholar contends that women and feminists have created an educated class that abhors men and is profoundly biased against them (somewhat plausible), and that this educated class has managed to cook the books as part of a broader conspiracy to convince people that men are violent and unruly (incredibly implausible). The first part of this claim might be edifying to debate; the second part, not so much, although obviously it should not be illegal to do so.

More extreme and obvious examples can be adduced without taxing one’s imagination. Consider, for example, the orthodox belief that murder is a heinous crime that deserves severe punishment. One can imagine a reasonable alternative to this. Perhaps murder, although reprehensible, is not quite so bad as is widely believed. And perhaps it would be edifying to debate those who contend that our society is too punitive toward murderers. That hardly means it would be edifying or even sensible to debate the position that murder is healthy and rejuvenating to a society. But not only would it not be edifying to have such a debate, but society is also almost certainly right to shame pro-murder arguments into silence. The same logic applies to orthodox beliefs about rape and robbery and racism, et cetera.

This might seem unfair to Mill. Surely, he does not intend to defend such extremes. But remember that he asserted (although he’s not always entirely consistent on this) that “unless the reasons [for free expression] are good for an extreme case, they are not good for any case.” Mill’s claim about extremes is obviously incorrect. Reasons good enough to justify a moderate case often fail to justify an extreme case because policies and norms are about costs and benefits. But his explicit claim is that the reasons that support easy or obvious cases of free speech must also support extreme cases, and one can thus judge his arguments according to his own stated criterion. This argument is not a persuasive defense of extreme forms of speech and dissent.

Furthermore, many community beliefs are value-based and not empirical in nature. It is unclear that disputation would or could lead a community to more accurate beliefs. Consider the age-of-consent laws again. I would not claim that they are entirely non-empirical because they are motivated by some real empirical concerns (e.g., about physical and mental maturity); but they are also based largely on moral values, not scientific data. And while some moral values are true or false to most people, others are less obviously so, and rest on shared commitment and deference more than they do on reason. Proponents of Mill would likely contend that beliefs and values that cannot withstand forceful debate are like ships that cannot withstand stormy seas: they are dangerous and deserve to be obliterated, not coddled. But that requires careful argument, not question-begging principles. And Mill does not provide those arguments.

An unchallenged orthodoxy, even if correct, becomes a meaningless catechism

Mill’s third major argument for free expression is, I think, his weakest and least persuasive. It is also unclear whether it is meant primarily as an empirical or a normative claim. As an empirical claim, it is obviously false. Most beliefs about the world are not etiolated in the absence of counterarguments. For example, most people believe that rape is wrong and have likely never seriously considered or debated the alternative hypothesis. This does not, however, lead to a kind of epistemic ennui; in fact, in some cases, the absence of a serious contradictory argument leads to enthusiastic certainty.

Moreover, the historical record suggests that open debate does not lead to certainty or ardent embrace of ideas, but rather to skepticism and sometimes to cynicism. Even beliefs as apparently obvious to ordinary people, such as that the external world exists, are doubted by trained philosophers and other educated people. Perhaps a proponent of Mill would counter that such beliefs, if they wither in the bright light of debate, do not deserve assent. Perhaps skepticism is the appropriate epistemic attitude. I can think of two responses.

First, even if this argument is correct, it undermines Mill’s position, since its premise was that the absence of serious debate would lead to the lamentable state in which “the meaning of the doctrine itself will be in danger of being lost, or enfeebled, and deprived of its vital effect on the character and conduct: the dogma becoming a mere formal profession, inefficacious for good, but cumbering the ground, and preventing the growth of any real and heartfelt conviction, from reason or personal experience.” One cannot bemoan weakly held dogmas because they are enervating while applauding a sophisticated skepticism that is equally enervating.

Second, this argument is almost certainly not entirely correct because it is premised on an erroneous view of human psychology. Most people do not know why they believe what they believe; and, more importantly, they will never know why they believe what they believe. And many laudable beliefs are therefore incredibly difficult to defend rationally because the reasons for them are hidden from the conscious mind. Take something as apparently obvious as opposing incest. Most people do not in fact know exactly why they think incest is wrong. And, as Jon Haidt argued in his book The Righteous Mind, when people are challenged about their views, they often become flustered and eventually dumbfounded, confessing, “I don’t know why. I just think it’s wrong!” Some even come to doubt their view that incest deserves moral opprobrium.

The proponent of Mill might reply that the fact people cannot offer good reasons for opposing incest suggests that at minimum certain forms of incest are not objectionable. Or the proponent might say that this is precisely why more debate is always better: “He who knows only his own side of the case,” wrote Mill, “knows little of that.” Thus, the disputation is necessary to promote a complete understanding of the case against incest. Whether or not these rejoinders are correct, however, does not address the point at issue—the empirical claim that unchallenged beliefs become ineffectual dogmas, at least when applied to an individual (as opposed to a community). In fact, the opposite is generally true: contradictory views and debate lead to skepticism and doubt, not certainty.

As a normative claim, Mill’s contention is slightly more plausible, especially if limited to specific communities such as scientists or philosophers or others whose jobs depend upon forwarding, defending, and debating ideas. Philosophers, for example, are often trained to protect apparently obvious and natural beliefs from hostile attacks. And therefore, philosophy professors often forward potentially offensive thought experiments or proposals about unconscious rape or nuclear holocaust to provoke and challenge their students.

But these debates are not designed to increase certainty in students. They are designed, instead, to increase the students’ abilities to debate effectively. And, in fact, I very much doubt that the reasons the students ultimately proffer commonsensical beliefs are the real reasons they hold or preserve those beliefs. As F.H. Bradley noted in his 1893 essay, Appearance and Reality, “Metaphysics is the finding of bad reasons for what we believe upon instinct, but to find these reasons is no less an instinct.” The average person believes that other minds exist because he is wired to do so, just as he is wired to enjoy a gorgeous sunset or cold water on a hot day. But the philosopher is not content with this instinctive response and strives to provide more logical or rationally defensible reasons. But the rationally defensible reason is not generally the causal reason for the belief.

It's important also to recognize that what is appropriate in a philosophy class at college is not necessarily appropriate on NBC News. During a lecture on human sexuality, a philosophy professor might challenge his students to argue against the proposition that having sex with non-human animals is perfectly moral, but if he made that same challenge on NBC News, the phone lines and email boxes of NBC executives would be flooded with outrage, disbelief, and indignation. (This hints at a more reasonable defense of free speech based on a kind of pluralism, which accepts greater free speech in some domains and lesser in other domains. In philosophy classrooms, free speech should be nearly absolute; on the nightly news, not so much.)

It is equally important to recognize that the value of debate does not lie in its ability to transform dead dogmas into living beliefs, but rather in its ability, in restricted enterprises, to lead to a more accurate view of the world. In this, Mill is probably correct, but only in a circumscribed way. In science, debate and dialectic often lead, through a painful and winding path, to the truth, but only because of other important institutional and ethical norms. In philosophy, debate and dialectic probably do not lead to the truth, but they can be mightily entertaining. (Ambrose Bierce hilariously defined philosophy as “A route of many roads leading from nowhere to nothing.”) Nevertheless, a cynic (or a sophist), might plausibly contend that dialectical prowess is valuable not because it allows one to home in on the truth, but because it allows one to manipulate other people; and therefore, that it is also valuable because it allows one to avoid being manipulated. A proponent of Mill might even concede this point, but this is surely less elevated and inspiring than the arguments in On Liberty.

At any rate, the major point here is that Mill’s claim about uncontested beliefs is wrong: They do not necessarily become tired, lazy, or empty dogmas. In many cases, in fact, unchallenged views are fervidly held, defended, and promoted. Somewhat ironically, one of the best ways to create ardent belief is not through debate, but through suppression and persecution, for groups that are persecuted often hold their dogmas with peculiar zeal. It is not philosophical dispute that creates strong beliefs, but tribalism.


Before concluding, I would like to adumbrate a defense of broad free speech, since, despite some reservations, I champion such a view, and I do not want the reader to think that my complaints against Mill are indicative of antipathy with his goals. In the domain of scientific discourse, expansive free speech is generally a positive good because it helps to guide us toward a more accurate view of the world. Science often contradicts or at least challenges sacred values and norms; and therefore, politicians, pundits, and even other scientists, will often be tempted to suppress hypotheses and ideas that are either true or an important part of an ongoing debate. This has the miserable effect of stifling scientific progress, which ultimately undermines society’s ability to address social, political, and technological problems. Mill’s arguments in a muted and more judicious form thus apply to science.

Perhaps more importantly, though, enforcement of suppression and censorship is costly and incentivizes undesirable behaviors. Suppose that one were to agree with a proponent of censorship: Some speech is dangerous (and if speech is effective, then surely it can be dangerous). It does not immediately follow that censorship is a wise policy. Most people agree, for example, that alcohol is dangerous. However, experience revealed that prohibiting alcohol was costlier than its benefits; consequently, prohibition is widely regarded as a failure, even though it did in fact reduce the consumption of alcohol (and alcohol-related disorders). Censorship, like prohibition, is costly. It angers people and creates divisions. It also encourages feigned outrage and indignation instead of engagement or tolerance, rewarding those who are the most angered and offended by silencing opposing speech. These are potentially steep costs and almost certainly are greater than teaching and promoting liberal attitudes about speech, especially in a large society with riotous intellectual and cultural diversity.

John Stuart Mill’s On Liberty is a passionate and eloquent defense of liberty which forwards bold arguments and a capacious conception of free speech. Many modern free speech advocates refer favorably to Mill’s work, and still rely on versions of his arguments to defend broad notions of free speech. However, like most classics, Mill’s work is almost certainly celebrated more than it is scrutinized, and most of its arguments, upon inspection, are weak, flawed, misleading, or only plausible in limited domains of discourse. It remains for those of us who have expansive views of speech to forward better and more plausible arguments.

PoliticsphilosophyFree Speech

Bo Winegard Twitter

Bo Winegard is an Associate Editor at Quillette. He received his PhD in social psychology from Florida State University under the tutelage of Roy Baumeister.