Will Wilkinson last week offered a thoughtful tweet storm about social penalties for making claims that are out of bounds:

Wilkinson insists that he favors free speech, in the sense that he believes that the government should not proscribe speech (outside of narrow categories, such as slander), but that all reasonable people exact social penalties for at least some speech. And indeed, while I consider myself as about as in favor of free speech as anyone, I can imagine some extreme statements that a dinner party guest might make (say, holocaust denialism or white supremacy) that would make me less likely to invite the guest to another party, in part because I am convinced that a person announcing such views is seeking to get a rise our of listeners, exhibits serious defects in reasoning ability, or has profound prejudices, or maybe all three.

The danger, though, is that once we accept that it is acceptable for there to be social penalties for making out-of-bounds claims, people who make claims that ought to be in bounds, maybe even claims that are correct, will be found to be out of bounds. Moreover, people will not make claims that they think plausibly might be out of bounds.

Indeed, Wilkinson confesses that he has “opinions I rarely share because I fear social blowback.” What are these opinions? Wilkinson doesn’t say. That is actually a bit surprising, because Wilkinson argues that “[w]e should just directly debate what claims ought to be unutterable by decent liberal people.” How are we to have this direct debate if we can’t report our own out-of-bounds opinions? Wilkinson appears to recognize this problem, acknowledging that “it’s hard to say that an opinion ought to be in-bounds without confessing that you hold an out-of-bounds opinion.” But he doesn’t offer a solution.

Maybe Wilkinson’s view is that one ought to be able to debate what opinions should be in bounds so long as one doesn’t advance the underlying opinions. But imagine the following claim: “I’m not a holocaust denier, but I think holocaust denial should be in bounds, because a lot of those photographs do look like they could have been faked.” It’s hard to imagine a world in which this claim receives substantially less opprobrium than the claim following the “because.” Indeed, the natural reaction of any listener would be to assume that the speaker is in fact a holocaust denier but is trying to avoid social opprobrium while still expressing denialist views, just as we may infer that someone who begins a sentence with “I’m not a racist, but…” probably believes the potentially racist sentiment that follows. And of course, one would receive even more opprobrium if one admitted, “I have a view that has been designated out of bounds, but I’m going to explain why I think it’s in bounds.”

If the debate about what is in bounds were limited to issues such as holocaust denialism and white supremacy, maybe it wouldn’t be worth worrying too much more about the problem. But Wilkinson strikes me as a reasonable, thoughtful person. I would be very surprised if he secretly were a denialist or a supremacist. But I know that there are mainstream opinions (like Steven Pinker’s) that are now the target of cancellation campaigns.

The knowledge that thoughtful people are self-censoring troubles me, not so much because it will lead me to censor myself, but because it makes it much harder for me and others to generate justifiable beliefs. Most of what any of us believes isn’t based on careful reviews of the literature. I believe in anthropogenic climate change and have even written about possible remedies for climate change, but I have not personally reviewed the models that predict global warming. My opinion is based on the declared opinions of others, who themselves may not have reviewed all the relevant models but may well be friends or friends of friends of people who have. I am, in other words, engaging in an exercise in social epistemology, trying to determine what is a justified true belief based on the announced beliefs of others.

But this exercise is a lot more difficult when one suspects that certain opinions are self-censored. If hypothetical climate scientists who have a view that differs from the consensus feel that they are better off staying quiet, then it is hard for an outsider to know whether the absence of such statements is because the climate change evidence is so strong or because there has been an information cascade. (The concern can push in the opposite direction as well. Because government climate scientists worry about stating their honest views, I would not place much epistemic weight on a government report about the state of climate science.) I still feel that I know enough about the culture of academia to determine with high confidence that climate change skepticism is largely unjustified. But I don’t have a very good answer to someone who, engaging in his or her own exercise in social epistemology, concludes that climate change is a hoax. I could tell this person that 97% of published papers that express a position on anthropogenic global warming conclude that it is occurring, but I don’t have a good answer to the objection that papers that say the opposite won’t get published and that scientists who claim such unorthodox views will harm their careers.

What I would like to be able to say to someone who raises a climate change hoax argument (or some other claim that I believe to be incorrect) is that the culture of academia encourages heterodoxy, and so where it is absent, a genuine consensus exists. I would like to be able to point to academics who raised heterodox positions (and by this, I mean something more radical and more likely to be wrong than anything Pinker would say, but probably not something so insupportable as holocaust denialism or white supremacy) and say, “That professor made a crazy argument, and received plenty of counterarguments but no public opprobrium.” But that is not possible today. Sure, academia is much better than most employers, because tenure remains a significant protection. Academic institutions censure without censoring, but that too can effectively silence those whose views are outside some range of permissible discourse, whether on the right or the left. And that makes it more difficult for observers, especially those outside the academy, to determine whether official socially acceptable positions are worthy of justified belief.

Academia should be a place where nothing is viewed as out of bounds, so that if everyone in academia seems to agree with proposition X, people who are outside academia but understand its culture believe with high confidence that X must be correct. I would not mind the occasional loony paper if the absence of condemnation for that paper improved the credibility of all the non-loony things that academics write. In my view, a culture that encourages debate is more likely to lead to wide social acceptance of propositions that are so clearly justified that they should not be controversial. This is an empirical claim, and I can’t be sure about this. Maybe allowing people to defend the indefensible makes it easier for outsiders to find at least one person who agrees with whatever they would like to believe. But I believe that creating institutions in which heterodox views are encouraged means that outsiders are more likely to trust orthodox views. If it really were not possible to tell the difference, orthodox views could be elicited through means such as surveys.

How can we make universities more tolerant of dissent? A start would be for universities to commit to the Chicago principles: “It is for the individual members of the University community, not for the University as an institution, to make those judgments for themselves, and to act on those judgments not by seeking to suppress speech, but by openly and vigorously contesting the ideas that they oppose.” So, no more letters from the university president disagreeing with views of a professor in a discipline the president may not know much of anything about. But universities can do much more to encourage open expression. Professors can encourage students to take positions that they don’t actually believe, both as an exercise for their own benefit and as a way of providing plausible deniability for others who take positions that they do believe. Professors too might be encouraged to write articles that take the best position that they can muster against what they actually believe. Sponsors of panels and workshops should always make sure to invite those with dissident views, or if no one is available to express such views, at least someone who will attempt to express disagreement to the best extent possible. We can be more confident in our own conclusions if we know that the arguments that we have heard are the best available on all sides of any debate.

Ideally, our commitment to free expression should extend beyond universities. Anyone trying to make good faith, thoughtful arguments, whether the speaker ultimately would endorse those arguments or not, should receive no social condemnation. If we are to condemn at all, it should be outside the spheres in which debate is vital, and what we should condemn even there should be not conclusions but incorrect premises and faulty logic, including ad hominem arguments or calls for cancellation. An approach that makes even Will Wilkinson thinks he should keep his mouth shut makes it too hard to determine what we are justified in believing outside our immediate domain of expertise.