Skip to content

Why Academic Freedom Matters

A frequently misunderstood argument in Mill’s On Liberty explains why academic freedom is essential in higher education.

In the midst of the 2020 protests for racial justice, I prepared to teach my annual July mini-course on free speech for incoming FLI (First-Generation and/or Low-Income) students at Amherst College. Almost all of these students are persons of color. My aim in this course is to show students how to analyze the reasoning in an important but difficult text from the history of philosophy, something for which most high schoolers have little training. The text I chose was John Stuart Mill’s famous defense of free speech in On Liberty. But I like to begin the course with a discussion of a specific free-speech controversy, so that, after reading On Liberty, students can reflect on whether this 19th century essay by a dead white male has something enlightening to say to us. Last summer, my idea was to have that initial discussion about Senator Tom Cotton’s recently published op-ed in the New York Times defending the federal government’s military response to the Black Lives Matter protests in Lafayette Square in Washington, D.C. After receiving heavy criticism from readers, the Times conducted a review, concluded that it should not have published the essay, issued a public apology, and eventually forced the resignation of James Bennet, the editor of the opinion page. Two Times opinion writers, Ross Douthat and Michelle Goldberg, wrote subsequent essays in the Times, one in favor (Goldberg) and one against (Douthat) these decisions. For the first day of class, I asked my students to read both essays along with the original Cotton op-ed, and write on the following question: Do you think the New York Times should have published Cotton’s essay? Why or why not? 

Of course, I was aware of some of the dangers of starting class with such an incendiary essay, but I also thought it was important for students to know right away that a good college education requires engaging with writings that might infuriate them, that they might even see as racist. My biggest fear was not that some students would accuse me of racism for assigning it or think that I endorsed Cotton’s position, nor was it that they would get in heated, uncivil disputes over the op-ed. What I feared most was that they would uniformly agree that Cotton’s essay should not have been published because they deemed his conclusion racist, and that Cotton’s incendiary rhetoric would so turn them off that they wouldn’t actually examine his argument. In fact, that is more or less what happened. Although one or two students thought Cotton’s essay should have been published even though it was racist, none of them talked about the argument he actually gave. 

Skip to the fall 2020 semester. Spurred by the summer protests for racial justice, and student demands for protection against racism on campus, the Amherst College faculty decided to revisit our statement on academic freedom. In our discussions, it was clear that most faculty saw our task as finding a balance between two potentially conflicting values: on the one hand, the value of having a campus free from the harms of racism; on the other hand, the value of faculty autonomy in choosing texts. Eventually, the faculty voted by a wide majority (90% in favor, 10% against, if I recall correctly) to amend the original statement on academic freedom to ensure that it did not permit speech that causes identity-based harm. Although I am still free to teach essays like the Tom Cotton op-ed, it was touch and go: there was discussion of forbidding the teaching of texts considered to be racist or morally repugnant in other ways. 

Mill presents an overlooked but powerful argument for freedom of speech that can also shed light on why academic freedom matters.

What I felt was missing in our faculty discussions was a clear articulation of the point of academic freedom, in particular at a small liberal arts college such as our own, something that I believe Mill can help us with. Mill presents an overlooked but powerful argument for freedom of speech that can also shed light on why academic freedom matters. His argument locates the point of free speech, not in a set of values such as freedom—which might, as our faculty decided, need to be curtailed by other values, such as respect and dignity—but in something more fundamental: the role of free speech in enabling us to rationally maintain our values and rationally act on them. Mill helps us to see the important, non-negotiable role that free speech plays in our lives as rational agents, in light of which censoring speech is self-defeating. Even if I am unable to convince you that Mill is right about this, I hope to at least entice you to read, or re-read, On Liberty, and to make up your own mind, as Mill would have wanted.

Mill’s Argument for Free Speech

At the beginning of Chapter Two of On Liberty, Mill outlines his case for free speech as follows: 

It is necessary to consider separately these two hypotheses, each of which has a distinct branch of the argument corresponding to it. We can never be sure that the opinion we are endeavouring to stifle is a false opinion; and if we were sure, stifling it would be an evil still.

Mill, p.14

It is during his discussion of the first of his hypotheses—“We can never be sure that the opinion we are endeavouring to stifle is false”—that Mill presents his ingenious argument for free speech. Mill begins his discussion of the first hypothesis this way: 

First: the opinion which it is attempted to suppress by authority may possibly be true. Those who desire to suppress it, of course deny its truth; but they are not infallible. They have no authority to decide the question for all mankind, and exclude every other person from the means of judging. To refuse a hearing to an opinion, because they are sure that it is false, is to assume that their certainty is the same thing as absolute certainty. All silencing of discussion is an assumption of infallibility. Its condemnation may be allowed to rest on this common argument, not the worse for being common.

Mill, pp.14-15

Before discussing this argument, we should pause to ask what Mill means by an “opinion”. Although Mill doesn’t explicitly discuss this matter, his central interest is clearly in political, social, scientific, and moral issues that are of public concern. It is a good question how to distinguish between topics that are of public concern and topics that aren’t, but we need not pursue it here. Mill’s argument should be of interest on any plausible way of drawing this line. 

When we silence an opinion, Mill says, our reason, at least in part, is that we are certain it is false. But in taking our certainty of its falsity as a reason for silencing an opinion, we are treating our subjective certainty as absolute certainty, which is tantamount to assuming we are infallible about the truth of our belief. But of course, we are not infallible beings; therefore we are not justified in silencing an opinion on the grounds of being certain it is false.

This “common argument” has been widely dismissed. It strikes most readers as resting on a premise that is obviously false, namely, that we must assume that we are infallible about the falsity of an opinion in order to suppress its expression. They reason as follows. Silencing an opinion is an action. Like all actions, it is based on a belief. But we generally do not require absolute certainty in our beliefs in order to act on them. Thus we do not need such certainty in order to be justified in performing the action of silencing an opinion. 

This objection relies on what looks to be an undeniable platitude about justified action, namely, that it requires justified belief but not infallibility or certainty. 

Just as my reasons for contributing to Oxfam, for example, can be its good reputation and track record, in which I believe but of which I am less than certain, so too Mill’s opponent can argue that they have good reasons for their decision to silence an opinion even though those reasons fall short of certainty. For instance, a college faculty or administration can claim it bans texts denying the existence of structural racism, not because it assumes it is infallible about the existence of structural racism, but because it has reasons for believing that structural racism is real and that allowing this important truth to be denied in the classroom is disrespectful to students of color. 

There is the greatest difference between presuming an opinion to be true, because, with every opportunity for contesting it, it has not been refuted, and assuming its truth for the purpose of not permitting refutation.

John Stuart Mill

This looks like a decisive rebuttal of Mill’s infallibility argument, but is it? In order to answer it, Mill needs to show either that justified action in general does indeed require infallibility—a fool’s errand—or that, unlike other actions, the specific act of silencing an opinion is unique in requiring infallibility. In this compressed passage, he takes the latter course: 

There is the greatest difference between presuming an opinion to be true, because, with every opportunity for contesting it, it has not been refuted, and assuming its truth for the purpose of not permitting its refutation. Complete liberty of contradicting and disproving our opinion, is the very condition which justifies us in assuming its truth for purposes of action; and on no other terms can a being with human faculties have any rational assurance of being right.

Mill, p.16

Difficult as this passage is to parse, it is worth the effort, because it contains Mill’s most interesting defense of free speech. In the first sentence, Mill points out that the assumption of an opinion’s falsity can be either the outcome of open debate in which that assumption has withstood all attempts to refute it or, instead, one’s reason for suppressing that very debate. In the final sentence, Mill concludes that it is only if one is willing to subject the assumption to open debate that one is justified in acting on it. How exactly does Mill arrive at this conclusion?

To understand Mill’s reasoning, we need to unearth the conception of justified belief that he is relying on. Mill argues that we are justified in our beliefs—that is, we can have rational assurance in assuming their truth—only if we are willing to confront reasons against their truth. Unlike infallible beings who can, as it were, see the truth directly, we can have rational assurance of the truth of our beliefs only by putting them through a process of justification. And, according to Mill, “on every subject on which difference of opinion is possible” this process of justification requires “a balance to be struck between two sets of conflicting reasons” (Mill, p. 30). To justify our beliefs, therefore, we must weigh the reasons for and against their truth and determine which reasons are stronger, or as Mill would put it, where the balance is to be struck. For instance, even if we have very good evidence that structural racism exists, unless we also consider the evidence that it does not exist, and weigh the total evidence to determine which proposition it supports, we will not be in a position to hold a justified opinion on the matter. 

Because silencing an opinion undercuts our ability to justify the assumption that the opinion is false—which is part of our justification for silencing it—the act of silencing an opinion undercuts its own justification.

Mill thinks we can engage in this weighing of reasons only if we are in an environment where the reasons both for and against our beliefs are available to us. By silencing the expression of opinions we deem false, we deprive ourselves of just such an environment. Our reasons for believing that an opinion is false must be weighed against any reasons for believing that the opinion is true, which requires express consideration of the opinion itself. Silencing the opinion thus undercuts our ability to justify our belief that it is false—which is part of our justification for silencing it—and so the act of silencing the opinion undercuts its own justification. Hence silencing an opinion can never be justified. When we act to silence white nationalists because we believe their views are false, we undercut our ability to justify our belief that their views are false, by preventing our reasons for believing so from being weighed against their reasons for holding those views, and thus we undermine our justification for the very act of silencing them. And this problem is not unique to the act of silencing white nationalist views. No matter the content of an opinion or the character of its proponent, the act of silencing it is self-undermining.

Is there a connection between Mill’s argument that censorship is self-undermining and his initial “common” argument that censors must assume that they are infallible? Mill himself does not spell out a connection, but here is the connection that he may have in mind. To think that we can be assured of the truth of our opinions without engaging in the ongoing process of weighing reasons for and against their truth—a process possible only in an environment of free speech—is to assume that we have the kind of direct access to the truth that only divine, infallible beings have. The thought that once we form an opinion we will never need to revise it in light of new evidence is just another instantiation of this assumption. This line of thought may explain why Mill continues to claim that, whether they know it or not, those who engage in censoring speech are in fact assuming that they are infallible.

The Necessity of Open-Mindedness

What Mill helps us to see is the indispensable role of open-mindedness in our lives as rational agents. Recall the end of Mill’s reply: “Complete liberty of contradicting and disproving our opinion, is the very condition which justifies us in assuming its truth for purposes of action; and on no other terms can a being with human faculties have any rational assurance of being right.” Mill does not say merely that one must test one’s beliefs against objections; he says that one must give others complete liberty to raise objections. Nor does he suggest that it is permissible to curtail this liberty after one has heard others’ objections; this would be incomplete liberty. In fact, two paragraphs later he says:

The beliefs which we have most warrant for, have no safeguard to rest on, but a standing invitation to the whole world to prove them unfounded. If the challenge is not accepted, or is accepted and the attempt fails, we are far enough from certainty still; but we have done the best that the existing state of human reason admits of; we have neglected nothing that could give the truth a chance of reaching us: if the lists are kept open, we may hope that if there be a better truth, it will be found when the human mind is capable of receiving it; and in the meantime we may rely on having attained such approach to truth, as is possible in our own day. This is the amount of certainty attainable by a fallible being, and this the sole way of attaining it.

Mill, p.18

Mill’s claim that we must give others “complete liberty” to disprove our opinion means that in order to be justified in acting on it, we must maintain a “standing invitation” to listen to new objections whenever they arise. According to this condition, we are not justified in holding an opinion merely in virtue of correctly weighing all the evidence that is available to us at some given time. Unless we are also willing to weigh new evidence in the future, we do not count, here and now, as being justified in acting on our opinion, no matter how well we have weighed the current evidence. To be willing to hear out such evidence is to be open-minded.

Philosophers—epistemologists in particular—will likely balk at the idea that a willingness to hear future evidence is relevant to whether one is justified in acting on an opinion. But we can see the role of open-mindedness in our judgments about justification if we examine the relation that future evidence has, not to one-off acts, but to the enactment of ongoing policies. It is the justification of a policy, after all, that is at issue in Mill’s argument: censors typically do not want to perform one act of silencing allegedly false, harmful opinions; they want to enact a policy of silencing such opinions over an indefinite period of time. As I hope will become obvious, the justification of ongoing policies requires that those who enact them be open-minded with respect to the opinions upon which the policies are based.

The connection between future evidence and policies may be easier to see if we first examine a case that doesn’t involve censorship. Consider whether a government is justified in enacting a policy of lowering taxes in order to spur the economy. The government may in fact have correctly weighed all the available evidence in forming its opinion that lowering taxes will spur the economy, but it has no guarantee that it has all the evidence there is or, more importantly, that there could be in the future. But if compelling new evidence against the policy is uncovered later, a refusal to modify or abandon the policy would be unjustified. And the government would be equally unjustified in trying to forestall that eventuality by announcing, when adopting the policy, that it will not consider any new evidence against it. A government that is unwilling to consider such future evidence is not justified in enacting the policy in the first place, no matter how good its reasons for the policy may be at the time of its adoption.

Why do I say that the government’s unwillingness to consider future evidence against its policy undermines its justification for adopting the policy rather than merely for continuing to carry it out in the future if evidence against it emerges? I say this because in adopting a policy, the government commits itself to implement it, not just at the moment of its initiation, but into the future. Such a future-oriented commitment requires that the government be prepared to assess future evidence against the policy whenever it arises and, depending on the results of that assessment, to modify or revoke the policy. The government’s unwillingness to consider such future evidence shows that it is failing to live up to this commitment. This is why a government that is close-minded is unjustified in acting on a policy, not only after evidence against the policy has been presented, but also in initiating the policy in the first place.

Let us now apply this lesson about the commitments of policies to the case of censorship. Following Mill, whose discussion is filled with religious examples, let us ask whether the government is justified in enacting a policy of silencing atheists on the basis of its belief in Christianity (but feel free to substitute a contemporary example if you like). At the time it first enacts the policy, the government may in fact have correctly weighed all the evidence that currently exists as to the truth of Christianity, but it has no guarantee that it has all the evidence that there could be in the future. If at some future date an atheist is able to slip through the censor’s net and present evidence against Christianity, the government would not be justified in refusing to consider that evidence and dogmatically sticking to its policy. And it obviously would make no sense for the government to justify sticking to its policy of silencing atheists by saying that, at the time it enacted the policy, the evidence supported the truth of Christianity. Furthermore, it would be absurd for the government to attempt to forestall objections by announcing in advance, when initiating the policy, that it will refuse to consider any new evidence against Christianity. Such announcements of close-mindedness by themselves would make the enactment of that policy unjustified.

Even if the government doesn’t announce that it is unwilling to hear new evidence in the future, its censorship policy will still be unjustified. By censoring atheists, it willfully deprives itself of the possibility of hearing such evidence and thus demonstrates that it is unwilling to hear it. Censorship is unique in this way. Other policies do not express an unwillingness to consider future evidence against them. For instance, the government can enact a policy of lowering taxes without thereby depriving itself of the chance to hear new evidence against this policy in the future. But the very act of instituting a policy of silencing atheists closes off the possibility of gaining new evidence against that very policy and thus makes the enactment of the policy unjustified. 

[I]f you are engaged in some purposive activity in the course of which speech happens to be produced, sooner or later you will come to a point when you decide that some forms of speech do not further but endanger that purpose.

Stanley Fish

Acting rationally requires that we be willing to subject our guiding beliefs to reflective scrutiny, which requires that we be open to challenges to the truth of those beliefs. If someone is unwilling to hear the evidence against a doctrine they believe, then they cling to it dogmatically, and dogmatism is anathema to rational, responsible belief: ostriches who bury their heads in the sand are not epistemic role models. The only way to avoid dogmatism in one’s beliefs and the actions based upon them is to be open to challenges to those beliefs. Censoring speech demonstrates that one is unwilling to hear such challenges, and thus the very act of censoring speech is inherently irrational.

But Isn’t Free Speech Itself Self-Undermining?

How does Mill’s argument for freedom of speech fare against the best argument for limitations on that freedom? Answering this question will bring out the depth and power of Mill’s argument and shed light on what he can tell us about the importance of academic freedom. The best argument for limits on speech is sharply expressed by Stanley Fish in his justly celebrated essay “There’s No Such Thing as Free Speech, and It’s a Good Thing, Too.” Just as Mill argues that censoring speech is rationally self-undermining, Fish argues that a policy of not censoring speech is rationally self-undermining. He argues, quite plausibly to my mind, that speech is valuable, not as an end in itself, but as a means to an end; and for whatever end one chooses, some types of speech will promote its achievement, whereas others will prevent its achievement. As he puts it, 

You assert, in short, because you give a damn, not about assertion—as if it were a value in and of itself—but about what your assertion is about. It may seem paradoxical, but free expression could only be a primary value if what you are valuing is the right to make noise; but if you are engaged in some purposive activity in the course of which speech happens to be produced, sooner or later you will come to a point when you decide that some forms of speech do not further but endanger that purpose.

Fish, p. 107

And when some form of speech endangers one’s purpose, “fidelity to the original values will demand acts of extirpation” (Fish, p. 103). Given that the value of speech is inherited from the value of the purpose to which it serves as a means, it would be rationally self-undermining, according to Fish, not to censor speech when it thwarts that purpose. To protect such speech would be to protect speech that undermines the very purpose that gives it whatever value it has. 

For Fish, a person who seeks to protect such speech is like a wealthy person who continues to accumulate money even though they have more money than they know what do with. We say that such a person fetishizes money: they treat an object that has merely instrumental value as though it is valuable for its own sake. So too, someone who advocates the protection of speech, even when that speech is inimical to their purposes, fetishizes speech: they treat an activity that has merely instrumental value as though it is valuable for its own sake. Viewed solely in terms of its value, championing unlimited or “free” speech is thus no more intelligible than making risky investments solely in order to have one more yacht or mansion that one will never use.    

To illustrate Fish’s argument, let us apply it to the case of a college that has the goal of training students to be anti-racists. This purpose certainly will give it reason to teach anti-racist texts. But this purpose also gives it reason to censor texts that will thwart this end, namely, texts arguing for policies that might turn some students into racists (e.g., texts arguing against affirmative action, or in favor of voter laws deemed to have a racially disproportionate impact). For a college to allow the expression of such views thus would be self-undermining; it would be putting at risk the very purpose—the fostering of anti-racist attitudes—for which the college values speech. 

This appears to be a compelling argument the college could offer for banning texts that express racist views. So what is wrong with it, by Mill’s lights? Put simply, by banning such texts, the college undercuts its own justification for instilling anti-racist attitudes. The college administrators’ efforts to instill such attitudes is based on their belief that racist views are false and that allowing such false views to proliferate would harm students of color. By deciding to censor the expression of such views, the college’s administrators demonstrate that they lack the virtue of open-mindedness necessary for them to be rationally justified in believing that such views are false and pernicious. But if they are unjustified in believing that such views are false and pernicious, they are unjustified in pursuing the goal of preventing students from holding such views, and thus unjustified in taking the means (i.e., censoring the expression of such views) to this goal. Here what is undermined is not a direct justification for silencing racist views—namely, that they are false and pernicious—but the college’s justification for seeking to extinguish those views in its students (because they are false and pernicious). Fish has introduced an ulterior end to which silencing false and pernicious views is a means. But the college’s rationale for taking this means is now undermined by its justification for pursuing that end.  

Admittedly, accepting Mill’s argument puts a college in a difficult position. By being rational and not censoring racist views, the college may fail in its mission to create an anti-racist campus culture and student body. But, Mill would say, these are just the risks that rational but fallible institutions must live with.

The Point of Academic Freedom

Might college administrators argue that racist views must be censored whether or not they are false, solely because their expression is offensive to its students of color? Mill himself considers just such an argument (Mill, pp. 18-19). He responds that in order for a judgment about the harmfulness of an opinion to be justified, one must allow free discussion about it; and that free discussion about the harmfulness of an opinion requires free discussion about its truth, because “the truth of an opinion is part of its utility.” Even if we rightly judge that the expression of an opinion may have some harmful effects, Mill agues, we cannot determine its overall harmfulness—its (dis)utility—without forming an opinion of its truth. Furthermore, I would add, once we decide to censor an opinion or idea solely because it causes offense, we leave the door open to censoring the expression of almost any idea, good or bad, true or false. As Fish himself says, “every idea is an incitement to somebody” (Fish, p. 106). 

[A]ssuming that the true opinion abides in the mind, but abides as a prejudice, a belief independent of, and proof against, argument—this is not the way in which truth ought to be held by a rational being. This is not knowing the truth.

John Stuart Mill

There is much more to say about Mill’s general response to arguments for censoring opinions even if they are true, but for our purposes, it may be better to focus on a consideration that is specific to the context of a liberal arts college. Notice that college administrators who censor views solely on the basis of their harmfulness would have to admit that by doing so they are condemning their students to holding unjustified beliefs about those views. This admission, however, is inconsistent with the mission of a liberal arts education. 

It is precisely because the mission of a college such as ours is to teach students how to be rational, responsible thinkers that I think Mill would find banning texts problematic. By banning certain texts from courses, we would turn the classroom into an arena of indoctrination and a breeding ground for dogmatism. To indoctrinate students is to persuade them to accept certain claims (e.g., that the military response to the protests in Lafayette Square was racist) by circumventing or manipulating their rational capacities. In such an environment, students do not come to believe a claim because they have weighed the best arguments for and against it; they believe it because they have listened only to reasons in favor of it. The message students receive when texts are banned is that it is permissible to believe certain doctrines without hearing (particular) arguments against them. But to accept a doctrine in this way is to hold it as a dogma, not befitting a responsible agent. As Mill nicely puts the point: 

[A]ssuming that the true opinion abides in the mind, but abides as a prejudice, a belief independent of, and proof against, argument—this is not the way in which truth ought to be held by a rational being. This is not knowing the truth. Truth, thus held, is but one superstition the more, accidentally clinging to the words which enunciate a truth.

Mill, p. 29

Ironically, if students are taught to dismiss arguments such as Tom Cotton’s because they are deemed racist by their professors or because of the inflammatory rhetoric with which they are expressed, students will hold onto their own anti-racist views as mere prejudices, even if those views are correct. If the Amherst faculty had voted to ban certain texts from being taught, we not only would have undermined our justification for doing so; we would have undermined one of the core parts of the college’s mission. Our most important role as a faculty body is to safeguard this mission. It is by losing sight of this part of our mission that we allowed ourselves to flirt with banning texts, and to accept a modification to our original statement on academic freedom that is bound to make it more difficult to carry out this mission. 

The calls to ban what are deemed racist texts and sanction professors who espouse or merely air views deemed racist is not that different from calls to ban communist texts and fire professors who espoused or merely aired communist views during much of the 20th century. The case for such bans and sanctions is no better now than it was then. As Alexander Meiklejohn, former Amherst College president and famous free-speech advocate, said in defense of allowing communists to teach,

[F]aith in freedom expresses our confidence that whenever, in the field of ideas, the advocates of freedom and the advocate of suppression meet in fair and unabridged discussion, freedom will win… . If that were not true, if the intellectual program of democracy could not hold its own in fair debate, then that program itself would require of us its own abandonment.

Alexander Meiklejohn

If our beliefs in democracy and racial equality are to be held rationally, not as mere prejudices, we must engage in such “fair and unabridged” discussion, even if doing so makes our views vulnerable to refutation. But for such discussion to actually benefit us, we must learn how to rationally weigh evidence and arguments; otherwise, no matter how much freedom of speech we have, we will not arrive at justified beliefs. And the best training ground of such rational deliberation should be in college and university classrooms, because teaching students how to think well is the fundamental purpose of a college education.

Editors’ Note: Readers wishing to pursuing these arguments further may want to consult the author’s paper “Why Censorship is Self-Undermining: John Stuart Mill’s Neglected Argument for Free Speech,” in Aristotelian Society Supplementary Volume 95:1, pp. 71-96.

Works Cited

Stanley Fish. “There’s No Such Thing as Free Speech, and it’s a Good Thing, Too,” in There’s No Such Thing as Free Speech, and it’s a Good Thing, Too (New York: Oxford University Press, 1994), pp. 102–119

Alexander Meiklejohn, “Should Communists be Allowed to Teach?”, The New York Times Magazine, March 27, 1949, p. 10

John Stuart Mill. On Liberty (Mineola, NY: Dover Publications, 2002 [1859])


Nishi Shah is a professor of philosophy at Amherst College.