Guest Post: Vinod Goel on Free Speech and Reason Responsiveness

A thoughtful commentary from a Cognitive Neuroscientist on the thorny issue of Free Speech.

I am part of a group of colleagues who regularly email each other about ideas and events associated with free speech and science and society. After my piece on Chat GPT that appeared here and in Quillette and in the National Post, my colleague Vinod Goel, who is a Professor of Cognitive Neuroscience at York University in Canada (the author of Reason and Less for MIT Press) wrote a thoughtful response to me about issues of artificial intelligence, education, and free speech. I asked him if he would write a guest post for Critical Mass and he kindly responded that he would.

I am attaching here a very carefully reasoned and comprehensive piece that he has entitled “Reason Responsiveness needs to be a pre-requisite for participation in free speech”. It raises a number of interesting and provocative questions that I am hoping will prompt reflection and discussion by Critical Mass readers. Please feel free to forward this piece broadly as I think the issues it touches on deserve wide dissemination. Vinod and I have engaged in a back and forth about some of them, and some of our discussions have made it into his piece, but the views expressed here are his. I thank him for taking the time to put his thoughts so carefully into writing for Critical Mass.

Americans are famously entitled to free speech. Free speech is, of course, not fully free. There are exceptions carved out for defamation, fraud, obscenity, child pornography, incitement, threats, and “fighting words.” But any citizen or corporation can insult or ridicule the president, arbitrarily proclaim that a vaccine causes autism, or insist that the 2020 presidential election was stolen by the ghost of Hugo Chavez, with impunity. The first strikes me as essential for a functioning democracy. The latter two give me pause.

The most famous argument for free speech appears in the second chapter of John Stuart Mill’s On Liberty. Here is a brief summary:

Truth is valuable. Any belief about the world will either be true, false, or partially true. There is no a priori way of knowing which is which. To determine the truth of a belief we must be free to evaluate it through evidence, reason, and discussion, and be at liberty to challenge and contradict it with differing beliefs in the marketplace of ideas, and be open to revision as necessary.

Suppression of beliefs is detrimental irrespective of their truth status. If the belief is true, suppressing it will hamper true descriptions of the world. If the belief is false, the assertion of false beliefs will lead to debate which in turn may lead to and deepen understanding of true beliefs. Finally, most of us can rarely see the whole truth, particularly when it comes to social, religious, and political discourse, so the only way of approaching it is through “reconciling and combining of opposites.” So whatever the epistemic status of a belief, suppressing it is harmful to the discovery of truth.

I think this is a good consequentialist argument. I would attribute many of the intellectual accomplishments of the Western world to at least some tolerance for free speech. But does this mean that to enjoy the benefits of free speech we need to accommodate beliefs about vaccines causing autism and the ghost of Hugo Chavez interfering in American elections?  I think what follows from Mill’s position is that we need to accommodate these and all beliefs, but we can deny a voice or platform to belief holders who are not reason responsive.

It is important to keep in mind that Mill was writing against the backdrop of (1) rationality, (2) shared worldviews, (3) acceptance of hierarchies, including knowledge hierarchies, and (4) editorial barriers to having your speech heard. In this piece I would like to take up the role of rationality in free speech and argue that the benefits of free speech only follow given rational actors. By rational I mean being sensitive to coherence relations or being “reason responsive.” For example, all of the Dragon fruits I have seen have been red on the outside. So I have come to form the belief that Dragon fruits are red. If I now see yellow Dragon fruits (or am told by a reliable witness that there are yellow Dragon fruits), being reason responsive requires that I adjust my belief that Dragon fruits can be either red or yellow. Later if I learn that what I took to be yellow Dragon fruits are actually a different variety of fruit, being reason responsive I would revise my beliefs again. If I fail to do so, in the absence of additional reason and evidence, I am not being reason responsive. I will argue that rationality or reason responsiveness of participants in a free speech community, at least locally with respect to the belief in question, is a necessary prerequisite for the success of Mill’s argument. Its absence provides reason to limit participation of certain individuals. I will only consider statements about how the world is, not how it ought to be.

I will make my case through three examples. The first two involve scientific/medical beliefs about ulcers and vaccines while the third is a political belief about the “stolen” 2020 American presidential election. With the first example I will argue that Mill’s marketplace of conflicting and diverging ideas is indeed absolutely essential for the discovery of certain true statements about the world, because the participants are rational or reason responsive to evidence. In the second example about vaccines, the same marketplace of ideas and evidence does not lead to belief revision among a section of the population, because they fail to be reason responsive. The third example illustrates the same point with a political belief. The bottom line is, in the absence of reason responsiveness, beliefs cannot contribute to our understanding of the world, irrespective of whether they are true, false or partially true.

Belief 1: Excess Stomach Acid Causes Peptic Ulcers

For decades the medical community was certain that peptic ulcers were caused by excess stomach acidity. There was widespread agreement among scientists and practitioners, and it was considered a closed issue. This conclusion was based on data and reason. In the 1980s two Australian scientists, Barry Marshall and Robin Warren, came along and resurrected an old discarded idea that ulcers were caused by bacteria, specifically, Helicobacter pylori. The bacteria theory of ulcers had been proposed multiple times over the previous 100 years but was repeatedly rejected by the medical community due to lack of conclusive evidence that bacteria could even survive in the acidic environment of the stomach. Therefore, many in the medical community initially ignored, even ridiculed Marshall and Warren and refused to believe their claims. The reputations and careers of individual scientists were at stake. The individuals who had developed, nursed, and defended the acid theory of ulcers were highly motivated to believe and advocate for it

But as long as individuals are committed to the scientific method and remain reason responsive, Mill’s marketplace of ideas should converge toward true beliefs. That is, among any group of scientists there will be many different beliefs and goals. There will always be some individuals who are not hampered by strong prior beliefs and vested interests in any particular theory and will be guided by the data (or commitment to an alternative theory). This plurality of beliefs and interests overcomes individual shortcomings and allows for convergence towards true descriptions. This is exactly what happened. Once the evidence was overwhelming, scientists did accept it and change beliefs. In 2005, Marshall and Warren shared the Nobel Prize in medicine for their conclusive demonstration that bacteria can survive in the stomach and cause ulcers. This changed the treatment of ulcers and improved the lives of millions of patients. This is an example of the successful revision of a long-held false belief (thought to be true) through reason responsive engagement with alternative beliefs and evidence generated by the scientific method.

Belief 2: The MMR Vaccine Causes Autism

There is a belief among a growing segment of the American in British population that the MMR vaccine causes autism. This is also a scientific medical question of fact, to be settled by evidence. While there has always been some residual public resistance to vaccination, this particular belief started with the publication of a scientific article by Andrew Wakefield in the respected peer-reviewed journal Lancet. In the article, he and his coauthors claimed a correlational link between MMR vaccination and autism in children. In subsequent public appearances, he made very strong causal claims. Other scientists were skeptical, but data are data. However, the results of the study could not be replicated by other labs, and it subsequently emerged that the paper may have been compromised. It eventually emerged that Wakefield was being funded by lawyers who were preparing a class-action lawsuit against the manufacturers of the MMR vaccine in the United Kingdom and may have “massaged” his data to fit the claim. The article was fully retracted by Lancet in 2010 when this information came to light. Wakefield was found guilty of serious professional misconduct and lost his medical license. Once the evidence was confirmed to be suspect, any beliefs based on it were quickly revised within the scientific community and among much of the general public, as demanded by reason responsiveness. It was agreed that there is no known credible link between the MMR vaccine and autism. All this is consistent with how Mill imagined free speech to work. But this is only half the story.

The other half the story has to do with the fact that a small but significant portion of the general public refused to revise and update their beliefs. They continue to loudly and widely proclaim that the MMR vaccine causes autism. This in itself is not problematic, and according to Mill, even essential to deepening our understanding of the subject matter, as long as they have credible reasons and evidence to support the belief. But no such credible evidence has been produced. There are anecdotal stories such as a mother reporting a diagnosis of autism in a child some weeks, months or years following immunization with the MMR vaccine, but no evidence employing accepted scientific methodology known to lead us towards truer descriptions of the world.

It is logically possible that the MMR vaccine causes autism. The belief itself is not the problem. The problem is that the assertion of this belief is not reason responsive to evidence for the contrary view, and it does not offer any coherent evidence and reasons in support of the belief. In such a case, it is difficult to see how the propagation of the belief can contribute to nudging us in the general direction of the truth. Also, notice that this is not an argument about expertise, though I think that is a separate relevant issue. It is about being reason responsive to the current data and having coherent reasons for rejecting certain findings and extrapolating other findings in particular directions.

Belief 3: The Stolen Election

The third example involves a socio-political belief. After the 2020 American presidential election, Trump and his supporters, and certain television networks, like Fox News, propagated the claim that there was massive orchestrated fraud and that Trump actually won the election. Some of the allegations were/are that Dominion voting machine software, allegedly owned by the family of the late President Hugo Chavez of Venezuela, switched millions of votes from Trump to Biden. It is certainly possible that there was massive fraud and Domain voting machines were the instrument to perpetuate it. Unlike in the above scientific cases, we cannot appeal to the scientific methodology to garner evidence one way or the other. But we do have a good methodology of testing the truth of nonscientific claims such as these. It is our legal system. The methodology of the courts is not infallible, but does a reasonably good job of uncovering facts.

Dozens of court challenges were mounted alleging massive voter fraud, and in each case the courts ruled that there was zero evidence of widespread fraud or malfeasance, and the election results as reported, were correct. Reason responsiveness requires that the findings of the legal proceedings lead to some appropriate revision of beliefs. But for Fox News and a large percentage of the MAGA faithful, no reason responsive revision occurred, but for very different reasons.

Given the information emerging from the Dominion lawsuit against Fox News, it is easy to understand what happened at Fox. They were actually reason responsive to the findings of the legal proceedings and privately did not believe the claims of widespread election fraud. However, they were also reason responsive to what their viewers wanted to hear and their financial bottom line, and made a conscious decision to encourage and repeat the false claims. This is perfectly rational and relatively uninteresting. It remains to be seen whether they will be held liable under the defamation laws for causing harm to Dominion.

What is much more interesting is that the MAGA faithful failed to be reason responsive to the court findings. Like the anti-vaccine group discussed above, they do exhibit rationality with respect to many of their beliefs (and will revise them in a reason responsive manner), but not with respect to these specific beliefs. This cannot be explained in terms of the standard accounts of heuristics and biases, sloppy reasoning, motivated reasoning, etc. I offer an explanation of it in terms of tethered rationality in Chapter 13 of Reason and Less: Pursuing Food, Sex, and Politics (The MIT Press, 2022). Given my current focus, I will not repeat it here.

Conclusion

In the case of both vaccine denial and election denial examples, the beliefs are unresponsive to reason. For Mill the whole point of free speech was approaching truer and truer descriptions of the world. This requires revision in response to reason and evidence. Without reason responsiveness there is no discussion, no give-and-take, no evaluation of data/evidence, no coherent conclusions; nothing to advance anyone’s knowledge of the truth or falsity of any proposition. Not only is there no possibility of advancement of knowledge in the absence of reason responsiveness, the possibility that the beliefs may cause harm remains ever present. One result of the anti-vaccine beliefs is declining vaccination rates in the United Kingdom and United States exposing children and other vulnerable community members to harm. In the case of the election fraud claims they disrupted the peaceful transfer of power. Harm to others has explicitly been recognized by Mill and other advocates of liberal democracy as a reason to limit individual rights. These are legitimate reasons to question whether participants who are not reason responsive with respect to their beliefs should have a seat at the table of free speech.

A pragmatic question is, what type of constraints might we impose to ensure platforms/voices are only offered to those who are reason responsive? A good starting point would be to reinstate the constraints that were widely present in Mills time, and indeed up until the development of the Internet: peer review and editorial filters. They have always been a part of publishing. They are still part of academic publishing. They are not perfect. (Like democracy, peer review is the worst system possible, except for all the others.)  I have several drawers full of manuscripts that have been rejected over the years for reasons that I do not agree with.  Upon rejection of my manuscript, I can be reason responsive to the reviewers concerns. Failing this, I can submit the manuscript to a different journal with a different editorial board and reviewers, etc. In this way peer review and editorial gateways not only allow for the marketplace of ideas, they give value and substance to the marketplace. A first useful step may be to adopt this system for all content platforms.

An effective way of doing this would be to simply repeal Section 230 of US Code 47 and treat all content providers as traditional publishers. They would then be responsible for the content that is posted on them, just as Oxford University Press and the New York Times are responsible for the content that they publish. This is all that would need to happen in terms of government intervention. Once private content hosting companies can be held responsible and sued (like traditional publishers) they would take the initiative to review and vet content. (I am only calling for a minimal vetting for reason responsiveness.) Who these reviewers are would depend on the nature of the platform. For a physics platform, you would presumably want physicists and not psychologists as reviewers. For news organizations, you would want them to enforce systems for fact checking and internal review built into traditional journalism (as we expect of the New York Times). For other platforms, nonspecialists from the general community would be appropriate reviewers (along the lines of a jury model). It may even be possible to have AI systems review certain content in the not-too-distant future. One can lay down better and worse practices, but who is selected, and the specific criteria, are up to the platform owners. Big Brother need not apply. Just hold all platforms legally responsible for their content postings (under current laws) and the free market will do the rest. This may reduce the amount of content posted, but that is just an additional bonus.

Some of my colleagues take exception to my consequentialist view and want to argue that free speech does not need any utilitarian justification; it is just an extension of our natural, absolute freedom over self. We are born free. In response to these views I would like to make the following brief points.

First, if one is unhappy with a consequentialist argument and prefers a deontological justification, then one needs an arbitrary rule or some metaphysical fact/principal to ground the claim for freedom. If the rule is arbitrary, then it is uninteresting. Those who have tried to derive such rules in less arbitrary ways ultimately end up grounding them in God's Grace or rationality. Philosophers from Aristotle to Kant have argued that humans are humans (or have “personhood”) because we are the rational animal. This is why I have certain inalienable rights that the chicken I ate for lunch does not. If you are going to ground your notion of personhood/humanity in rationality, then it seems to me that you should also endorse the reason responsive argument that I am offering.

Second, suppose that I just grant the principal of “an inalienable right of freedom." Unless I live on a deserted island, how far does this freedom get me? For example, suppose I rent the apartment above you and like to sing, stomp, and play loud music whenever it pleases me, usually at 3 AM. As soon as I exercise my freedom to do so, I am bumping up against your right to quietly enjoy your home. Unless my freedom trumps your freedom, or we settle through violence, there must be some socially negotiated constraints on our respective freedoms. One widely accepted constraint is the principle of minimizing harm to others. But this is a potentially dangerous, thorny issue when the rubber hits the road.

So it seems to me that whether you approach freedom of speech via consequentialism or deontology, there are limits. Slogans such as “born free” have a nice ring and fit on placards, but the notion of absolute freedom is an unhelpful fiction. It only applies when I am the sole occupant of a deserted island, and then it is unnecessary. The real work that needs to be done is laying out ground rules for social interaction. If I am not reason responsive to these rules my freedoms will need to be curtailed to preserve your rights.

Returning to free speech, privately one should be able to think and believe whatever one wants. However, before one is given a platform to widely disseminate descriptive beliefs, one’s adherence to those beliefs must be reason responsive (and not result in harm to others).

-Vinod Goel