Mill’s defense of free speech rests on the idea of a competition amongst arguments and ideas. We should avoid censoring speech, Mill argues, because we might find that unpopular ideas are in fact true, and even if they are wrong, they may help us come up with better arguments for better ideas. This is a compelling model of speech, both because it offers an interesting procedural account for how we might expand the stock of knowledge and arguments, and it provides an optimistic account of progress through opposition.
On this model, contentious speech is valuable, whether right or wrong, because it demands a critical response, and in making those critical responses, we come to better understand the contours of the truth, the circumstances under which the truth holds, and what makes a true belief true. This helps us not merely have true beliefs, but have knowledge. Free speech is then not merely valuable for improving the stock of knowledge, but for improving the deliberative capacities of the agents who participate in the marketplace of ideas.
Mill’s defense of free speech is powerful. However, the picture he offers is incomplete. Most notably, Mill fails to consider the possibility that the participants in the marketplace of ideas may well be a function of the content of those ideas. I suggest that there are at least two mechanisms that can cause a reduction in participation: a silencing effect such as that outlined by Langton (1993), and a discouraging effect. The first mechanism – one of silencing one’s opponents by means of undermining their epistemic authority – was first articulated in the case of pornography by Langton (1993), and made much more general by Fricker (2007). This mechanism works by using speech to de-legitimize a group of potential contributors to the marketplace of ideas. Their speech becomes discounted or altogether dismissed, despite their interest in countering the wrong speech from the original speaker.
The second mechanism – discouraging one’s opponents – relies on participants exiting the epistemic community of their own accord. In this set of cases, agents exit the epistemic community when they decide that the personal costs of participating outweigh the expected social gains that their participation would generate. This relies on an important aspect of Mill’s account of free speech. Recall that he agreed that wrong speech can generate unhappiness, but that this would be outweighed by the eventual triumph of the truth and the gains to our understanding of how the arguments for the truth work. While this long-run argument is surely correct, it does not account for incentives in the short run. If it’s too emotionally taxing, exhausting, or potentially dangerous to participate in a debate, agents will reasonably exit the debate even if they could offer useful arguments to counter the relevant speech.
Putting these two mechanisms together, we can see that some speech can serve to reduce the overall amount of speech, by imposing epistemic burdens on others who might have otherwise offered a reply.
If we consider instances of campus speech, this sort of concern is widespread across the political spectrum. Conservatives, for instance, have claimed that the hostility to their views is such that they feel uncomfortable raising them for fear of being labeled a racist or a sexist or a homophobe. Likewise, various minority groups have voiced concerns that it is simply too exhausting to rebut what they feel is a steady stream of animus against their social identity and mischaracterization of their views. Many parties on campus find themselves feeling like they are unfairly burdened by the speech of others, making it harder for them to respond.
The discussion around “safe spaces” can indeed be understood in terms of this view of the personal costs of speech. While the Millian idea of the marketplace of ideas is clearly a social good, individuals have good reason to not want to argue all day long. A perfectly reasonable interpretation of a safe space is just an environment where one is not called on to constantly defend their beliefs, and instead be in an environment with like-minded individuals at least some of the time. Those with minority views can feel this need more acutely simply because there are in a more hostile speech environment. People whose views are more in line with the majority in their community don’t feel the same need for a safe space, simply because the relative burden of speech is lower for them in general.
If we take these mechanisms for exit and entry seriously, this suggests that Mill’s argument for free speech needs to be expanded upon. Maximizing speech is no longer merely minimizing the regulations that we place on the content of speech, but instead is a function both of the amount of speech any given participant makes, and the total number of participants. Most importantly, however, if we wish the marketplace of ideas to function in the way that Mill argues, we have reason to worry at least about the composition of speech, and therefore the composition of the participants. After all, if we want to actually maximize speech, we ought to maximize the productive give-and-take that Mill envisions.
Our challenge, then, is to buttress the Millian idea of speech drawing out more speech by ensuring that we maintain a diverse set of participants in the marketplace of ideas. As we have already seen, one approach is by means of ensuring adequate representation of various positions and perspectives, such that the personal cost of participating is low enough to help draw people into the discussion. Another way of doing so is by attempting to balance the burdens of speech.
One approach would be to simply ban any speech that silences or discourages participation in the marketplace of ideas. On a number of college campuses, we see demands for versions of this approach – that speakers should be denied a platform to speak they have previously expressed views that (liberal) protesters deem to be silencing or discouraging. Likewise, we have seen demands for some courses or professors be removed from campuses for teaching material that conservatives find denigrating. I believe this approach is ill-advised, for two reasons. First, and most importantly, it is simply too powerful a weapon to wield against an epistemic (or political) opponent. This would lead to victory by default in the marketplace of ideas. Second, it might serve to discourage participation more than encourage it, by raising the costs for reasonable participants who mean well but hold a minority view.
What, then, can we do in place of the fairly straightforward approach of banning burdensome speech? One option is for campuses to try and balance how much burden any given group is willing to accept against how much they can potentially burden others. In this way, we can create a bargaining situation that can be assessed in two ways: first, by the volume of speech generated, so as to work to satisfy our commitment to a wide-ranging commitment to free speech, and second, by the degree to which different participants in speech have distributed the very real burdens and benefits that can be generated from free speech. We should aim for the largest marketplace of ideas that is consistent with relatively equal burden-sharing.
One might protest that this may seem to favor non-speech considerations over the widest extent of speech possible, but I think that this argument is wrong. What I propose is meant to capture the conditions under which contentious speech draws out more speech, rather than less. If we allow for disproportionate burdens in the population, that primes the conditions for speech that can silence, and as the composition of speakers moves away from a diverse representation of a variety of views, it is more likely to increase the costs of participation for epistemic minorities. If we instead balance these burdens, we can maintain adequate representation, minimize the conditions under which a silencing mechanism could work, and increase the scope of participation.
This suggests that our epistemic interests in fact line up rather well with our broader ethical interests in providing conditions of mutual respect. For contentious speech to work to invite more speech, it must be the case that there is an expectation that the response will be considered by the wider epistemic community. Without such expectations, we lack the preconditions for productive epistemic exchange.
What might this look like in practice? In the campus context, it might require that campus communities do more to coordinate speakers across distinct epistemic groups, such that the community finds the total composition of those speakers to be broadly balanced, rather than fixating on any one speaker. Likewise, there may need to be more emphasis on greater epistemic diversity in other areas of campus life, including in the classroom. The metric for success is not whether every last opportunity for speech is seized, but rather whether the ecosystem for productive contentious speech is maintained.