The discussion over no-platforming is often presented as a debate between proponents of free speech, who think that the only appropriate response to bad speech is more speech, and those who think that speech can be harmful. I think this way of framing the debate is only half-right. Advocates of open speech emphasize evidence, but they overlook the ways in which the provision of a platform itself provides evidence.
No-platforming is when a person is prevented from contributing to a public debate, either through policy or protest, on the grounds that their beliefs are dangerous or unacceptable. Open-speech advocates highlight what we might call first-order evidence: evidence for and against the arguments the speakers make. But they overlook higher-order evidence.
Higher-order evidence is evidence about how beliefs were formed. We often moderate our confidence in our beliefs in light of higher-order evidence. For instance, you might find the arguments in favor of booking plane tickets to Las Vegas right now compelling, but hesitate in light of the fact that you're really drunk. The fact that you're drunk is higher-order evidence: It's evidence that you might not be processing first-order evidence well. Maybe the case for going to Las Vegas won't seem quite so compelling in the morning.
Higher-order evidence has been the focus of a great deal of discussion in contemporary epistemology, under the heading of the epistemic significance of disagreement. Here's a familiar example: You and a friend have decided to split the bill 50/50 for your meal. Each of you separately adds up all the items you both had, using the prices listed on the menu (the bill is taking forever to arrive), and divide the total by two. You conclude that you each owe $34, but she thinks the right amount is $36. You know that you're more or less equally reliable at mental arithmetic. In this case, you have all the relevant first-order evidence easily at hand, but when you discover the disagreement, you should reduce confidence in your answer. Having discovered that your friend disagrees, you should recheck the calculation. At least one of you has made a mistake, and you have no more reason to think it's her than you. So you should lower your confidence in your belief.
An invitation to speak at a university campus, a prestigious event, or to write an opinion piece for a newspaper provides (prima facie) higher-order evidence. It is evidence that the speaker is credible, that she has an opinion deserving a respectful hearing. It typically certifies expertise, and expertise is higher-order evidence that the person's opinion should be given particular weight. Consider a case such as the restaurant bill example above, but now the math involved is difficult, and the dispute pits you against a math PhD; it's clear you should think it is much more likely that you have made a mistake than that she has, and you should defer to her.
Epistemologists have pointed out that we can sometimes hold fast to our beliefs in the face of disagreement, because the best explanation of why she disagrees with me is that she isn't competent about this matter — or she's joking or trolling me. An invitation to speak at a credible venue is evidence that the person is competent and sincere. It therefore provides weighty higher-order evidence.
An invitation to speak confers credibility because it selects someone from among the large crowd of potential speakers as the person whom we should hear. The credibility stems both from the fact of selection and from the prestige of the venue where they are selected to speak. By contrast, self-selection confers no credibility at all. If I jump on a soapbox at Hyde Park corner, I have self-selected, and no extra credibility is conferred on me. (If anything, speaking in this kind of venue confers negative higher-order evidence: People might think I'm a crank and take me less seriously.) The strength of the higher-order evidence conferred by an invitation depends on the prestige of the venue — being invited to give a TED talk confers a great deal of credibility, since the organization is so prestigious — and facts about how the selection procedure is taken to work.
To see how this plays out in practice, consider the controversy over the documentary Root Cause (2019), which peddles the discredited view that root-canal treatments cause cancer. The documentary is available on Netflix and Amazon. In an interview with The Guardian recently, media scholar Erica Austin of Washington State University said that unlike YouTube, where content is user-generated, being available on those sites confers credibility on the documentary. Being on YouTube is like getting up on my soapbox: No extra credibility is conferred on me by self-selection. But being selected for Netflix or Amazon confers some credibility. Of course, Netflix and Amazon are not very selective, so the credibility conferred is quite minimal. Being invited to speak at a university confers much more credibility.
Higher-order evidence is genuine evidence. It is rational to respond to higher-order evidence by moderating our confidence in our beliefs, sometimes even abandoning them altogether (as when I find myself disagreeing with someone who has much more expertise than me). Advocates of no-platforming therefore can cite epistemic considerations in favor of their view. They can reasonably argue that inviting someone to give arguments that are bad or false generates misleading evidence, and we should avoid generating misleading evidence. If someone is likely to speak in favor of a view we know to be false, we have grounds to no-platform them, because we know that providing them with a platform by itself provides higher-order evidence in favor of that view.
Importantly, higher-order evidence is extremely difficult to rebut. If my university gives a platform to a climate-change skeptic, it provides higher-order evidence in favor of her view. That higher-order evidence is not rebutted by the university inviting another speaker later to 'balance' her, or if she is subject to a devastating response from the floor. We can rebut her claim that global warming isn't occurring, but we cannot rebut her claim that the invitation certifies her expertise.
Of course, the invitation is only prima facie higher-order evidence. If we discover that she was invited only because she made a big donation to the school, or because she is the dean's cousin, that evidence is considerably weakened. It is this kind of consideration — not rational argument — that rebuts higher-order evidence. We rebut higher-order evidence using approaches that many open-speech arguments deplore because they don't address first-order evidence. An ad hominem attack (he's funded by the oil industry; he's a racist) or attacks on the credibility of those who provided him with a platform do not address his arguments, but they are often appropriate responses to higher-order evidence.
Free speech is genuinely valuable. Knowledge is produced through argument and debate, and good societies allow the expression of dissent. There are costs to suppressing speech. But the fact that the provision of a platform provides higher-order evidence in favor of a view entails that there are epistemic considerations on both sides of this debate. When we have good reason to think that the position advocated by a potential speaker is wrong, we have an epistemic reason in favor of no-platforming: We can be confident that providing her with a platform will produce evidence in favor of her views that it is very difficult to rebut (and which can't be rebutted by argument). Sometimes, at least, this consideration will be weighty enough to justify refusing to provide speakers with a platform.