1.1 ‘Crazy’ and ‘Dangerous’ Ideas

Counter-deniers typically take it for granted that science denial is a concerning phenomenon, and they usually make little or no effort to explain why it is so. However, before we can assess the validity of the counter-deniers’ concerns about science denial, we must first understand them. In this section, I use the counter-deniers’ scattered remarks on the topic to try to bring into sharper focus the exact nature of their concerns.

One of the clearest expressions of the counter-deniers’ concerns can be found in the opening lines of philosophers Steven Nadler and Lawrence Shapiro’s book When Bad Thinking Happens to Good People (which are also dramatically emblazoned in large orange letters on its red dustcover):

Like Obama, Nadler and Shapiro are not exclusively concerned with science denial – they are concerned with the broader epistemological crisis that is allegedly engulfing liberal democracies.Footnote ¹⁰ And they too, like Obama, consider science denial one of the most serious symptoms of that crisis.Footnote ¹¹

Nadler and Shapiro’s opening remarks (along with many other scattered remarks by counter-deniers)Footnote ¹² suggest that their concerns about science denial can be classified into two broad categories – epistemic concerns and practical concerns. Epistemic concerns include the concern that some people embrace ‘crazy’ ideas. Practical concerns include the concern that some of those ‘crazy’ ideas might be ‘even dangerous’.Footnote ¹³ The rest of this section examines these concerns more closely. Section 1.2 focuses on the counter-deniers’ practical concerns about science denial, Section 1.3 turns to their epistemic concerns about it, and Section 1.4 argues that the practical concerns should be given priority over the epistemic ones.

I should note that, while I have doubts about many of the factual and normative assumptions that underlie the concerns that motivate counter-deniers, the main goal of this section is not to assess the validity of those concerns but to better understand them, as it is important to develop a clearer understanding of the counter-deniers’ views before evaluating or criticizing them.

1.2 Practical Concerns

The counter-deniers’ practical concerns about science denial focus on the (alleged)Footnote ¹⁴ practical consequences of science denial. The broadest practical concern is that science denial might be (directly)Footnote ¹⁵ (ex ante)Footnote ¹⁶ practically harmful – that is, it might directly contribute to science denier’s making decisions that have practically harmful consequences. McIntyre, for one, claims that science denial enables ‘hundreds of thousands of people to refuse to vaccinate their kids, politicians to refuse to take action on climate change, and gun-toting protestors to parade during a pandemic’ (McIntyre Reference McIntyre2021, 28).Footnote ¹⁷

It is helpful to distinguish three distinct levels at which science denial might be practically harmful – personal, interpersonal, and societal. Science denial is personally practically harmful to the extent that it directly contributes to science deniers’ making decisions that have practically harmful consequences for them. For example, during the COVID-19 pandemic, many outspoken COVID sceptics died of the virus after refusing to take basic precautions against it.Footnote ¹⁸ While, in the absence of systematic empirical studies, we cannot be confident that their scepticism about COVID-19 played a causal role in their deaths, it is plausible to assume that people who underestimated the seriousness of COVID-19 were less likely to take precautions against it and more likely to contract the disease than people who took the disease seriously. An even clearer case of the practical harmfulness of science denial at the personal level might be the case of HIV/AIDS denial, as there is evidence that it results in many HIV-infected patients refusing antiretroviral therapies that significantly reduce mortality and morbidity due to HIV/AIDS.Footnote ¹⁹

Science denial is practically harmful at the interpersonal and societal levels to the extent that it directly contributes to science deniers’ making decisions that have practically harmful consequences for other people or, more generally, for the rest of society. For example, a parent’s refusal to have their children inoculated with the MMR vaccine has harmful practical consequences for their children (as it increases their chances of contracting measles, mumps, or rubella), for those who come in contact with their children (especially those who are too young to be vaccinated or who cannot be vaccinated for medical reasons), and for society at large (as it contributes to undermining its herd immunity against those diseases).Footnote ²⁰ Again, one of the clearest cases of interpersonally and socially harmful science denial might be HIV/AIDS denial, as there is evidence that HIV-infected HIV/AIDS deniers are more likely to have unprotected sex with HIV-uninfected partners, which is both practically harmful to their sexual partners and practically harmful for society, as it contributes to the spreading of HIV.Footnote ²¹

Lastly, it is useful to single out a specific kind of practical concern at the societal level – namely, the concern that science denial that may have practically harmful consequences at the political level. The main concern here is that science denial can have detrimental effects on both political debate and policymaking.Footnote ²² For example, voters who do not believe in anthropogenic climate change are more likely to support political candidates who oppose policies aimed at mitigating anthropogenic climate change.Footnote ²³

Another political concern is that science denial can reduce voluntary compliance with beneficial recommendations or regulations. For example, vaccine-hesitant parents often circumvent vaccine mandates by claiming spurious religious exemptions.Footnote ²⁴ Even if it is in principle possible for governments to coerce citizens into complying with their directives (by, e.g., issuing stricter childhood vaccine mandates), it seems to be neither desirable nor feasible for a liberal democracy to resort too extensively to coercive measures.

1.3 Epistemic Concerns

Let me now turn to counter-deniers’ epistemic concerns about science denial. Epistemic concerns can be sorted into two broad categories – deontic and consequentialist. Consequentialist epistemic concerns are concerns about the epistemically harmful consequences of science denial.Footnote ²⁵ Again, it is useful to distinguish three levels at which science denial can be epistemically harmful – personal, interpersonal, and societal. The main consequentialist epistemic concern at the personal level is that science denial can be epistemically harmful for the science deniers themselves, as those who embrace one form of science denial might become more epistemically vulnerable to other forms of science denial or, more generally, to unjustified beliefs.Footnote ²⁶ The hypothesis is that, for example, becoming a flat-Earther might increase one’s chances of becoming an anti-vaxxer, a climate-change denier, or a 9/11 truther. Several mechanisms may underpin the (alleged) increased epistemic vulnerability of science deniers. For example, it might be due to the interconnectedness of our beliefs, or to its effect on our epistemic character.

The main consequentialist epistemic concern at the interpersonal and social level is that science denial can cause an epistemic contagion – that is, the further spread of one form of science denial from one person to another and, more generally, through society. The hypothesis is that exposure science deniers’ beliefs could convert others to science denial. The most obvious potential route of epistemic contagion is the direct route, which consists in one science denier persuading others to become science deniers.

Finally, here too, it is appropriate to distinguish one subspecies of epistemically harmful consequences at the social level – that is, those that are epistemically harmful at the political level. For one thing, science denial might contribute to undermining the kind of healthy public debate on which the proper functioning of liberal democracies depends by depriving the debaters of a shared ground of factual beliefs.Footnote ²⁷ For another, science denial might affect education policy. For example, in the United States, creationists have engaged in various efforts to have creationism taught alongside evolutionary theory in biology classes.Footnote ²⁸

Deontic epistemic concerns about science denial focus on the epistemic conduct of science deniers. One possible deontic epistemic concern is that science deniers are epistemically incompetent – they do not have ‘the capacity to distinguish what’s true from what’s false’ (as Obama put it in his interview with The Atlantic). However, the most common deontic epistemic concern seems to be that science deniers are epistemically irresponsible.Footnote ²⁹ This concern finds particularly clear and forceful expression in the following passage from Nadler and Shapiro:

The concern seems to be that phenomena such as science denial result from an epistemically blameworthy kind of epistemic irresponsibility rooted in some kind of epistemic vice. While Nadler and Shapiro call the alleged epistemic vice ‘epistemic stubbornness’, others seem to think of it as something closer to what Quassim Cassam (Reference Cassam2019) calls ‘epistemic insouciance’, which is an indifference to (or lack of concern for) the truth.

These deontic epistemic concerns converge with the post-truth diagnosis, which attributes the epistemic malaise afflicting liberal democracies to a lack of interest in (or concern for) the truth (or any other epistemic goals) on the part of a growing number of their citizens.Footnote ³¹ Since science denial is often considered one of the key symptoms of that malaise, science denial can be regarded as a crucial test case for a post-truth diagnosis. If it turns out that, as many counter-deniers seem to believe, science denial is best explained in terms of science deniers’ lack of interest in (or concern for) the truth, then this would support a post-truth diagnosis of the epistemic malaise afflicting liberal democracies.

While it is too early to try to assess the validity of the counter-deniers’ concerns, a few remarks about the significance and the legitimacy of deontic epistemic concerns are in order. The first remark is that deontic epistemic concerns are not concerns about science denial per se – they are concerns about the deeper epistemological problems that (supposedly) underlie phenomena such science denial (such as our fellow citizens’ (alleged) epistemic incompetence or their (alleged) lack of interest in (or concern for) the truth). If so, then science denial is merely a symptom of deeper problems.

The second remark concerns the political legitimacy of questioning epistemic competence in liberal democracies. Historically, accusations of epistemic incompetence have served to marginalize and disenfranchise already disadvantaged social groups (including women, ethnic and racial minorities, people with disabilities, and the poor).Footnote ³² Even today, claims about the alleged epistemic shortcomings of our fellow citizens are often used to undermine the ideals of democracy and challenge the political authority of democratic governments.Footnote ³³ For these reasons, in a liberal democracy, citizens must operate on the presumption that our fellow citizens are epistemically competent unless proven otherwise.Footnote ³⁴ Moreover, as liberal democracies become more polarized, we must exercise particular caution in levelling accusations (or even insinuations) of epistemic incompetence or poor epistemic character, as such charges can all too easily be used to dismiss without due considerations the sincerely held beliefs and genuine concerns of those who disagree with us. Of course, this does not mean that we should naïvely assume that all our fellow citizens invariably act in good faith. For instance, it is undeniable that some actors deliberately spread disinformation on scientific topics to advance their own social, economic, or political agendas.Footnote ³⁵ However, there is a crucial distinction between challenging the motives of those who act in bad faith and impugning the epistemic competence or the epistemic character of those who act in good faith.

The third remark is that it is unclear whether deontic epistemic concerns are compatible with one of the pillars of liberal democracy – the principle of epistemic tolerance, which holds that everyone is entitled to their own beliefs, no matter how wrong or bizarre they may seem to others. To an atheist, the religious beliefs of Christians, Muslims, or Hindus might seem no less bizarre or unfounded than the beliefs of flat-Earthers seem to counter-deniers. However, in a liberal democracy, everyone is entitled to their own beliefs, no matter how bizarre or unfounded those beliefs might seem to some of their fellow citizens, and this seems to apply as much to the sincerely held beliefs of flat-Earthers as to those of Mormons or Jehovah’s Witnesses. Of course, this does not mean that citizens of a liberal democracy must refrain from trying to rationally persuade other citizens to change their mind. However, rational persuasion differs significantly from publicly impugning someone’s epistemic competence or responsibility.Footnote ³⁶

Counter-deniers may reply that political rights come with a host of associated responsibilities and that, in the case of freedom of opinion, the relevant responsibilities are primarily epistemic responsibilities.Footnote ³⁷ On this view, those who fail to fulfil their epistemic responsibilities essentially forfeit their right to their own opinion. One problem with this view, however, is that it seems to uphold epistemic tolerance in theory while undermining it in practice. What would prevent an atheist (or theist) from arguing those who disagree with them forfeit their right to their own religious beliefs because they fail to fulfil their epistemic responsibilities? Moreover, even if freedom of opinion were conditional on epistemic responsibility, it would still be incumbent on counter-deniers to substantiate their claim that science deniers do, in fact, fail to fulfil their epistemic responsibilities. As I argue in this Element, that task is much harder than counter-deniers typically seem to think.

1.4 On the Primacy of Practical Concerns

While this brief survey is unlikely to be exhaustive, it provides a preliminary taxonomy of the concerns that seem to animate counter-deniers, which, in turn, might help bring into sharper focus the nature of the phenomenon that primarily concerns them. While I am not persuaded of the validity or legitimacy of most of these concerns, I want to set my worries aside for the moment. Instead, I want to argue that, even if all counter-deniers’ concerns were valid and legitimate, they would not be all equally important or urgent – the practical concerns would have priority over the epistemic ones. In fact, I would argue that, ultimately, it is appropriate to be concerned about science denial only to the extent that it is practically harmful.

Let me start with what I have called deontic epistemic concerns. Deontic epistemic concerns include the concern that science denial might be a symptom of a deeper epistemic ill, such as the (alleged) epistemic incompetence of science deniers or their lack of concern for the truth. Those who doubt the primacy of the practical concerns might note that deontic epistemic concerns have nothing to do with the practical consequences of science denial, as they are explicitly purely epistemic concerns. On closer scrutiny, however, this claim does not seem correct.

To see why, suppose, for example, that all forms of science denial were completely practically harmless.Footnote ³⁸ Would deontic epistemic concerns still be legitimate? Counter-deniers might insist that, even if all forms of science denial were completely practically harmless, science deniers would still be epistemically blameworthy for their epistemically unjustified beliefs. They might insist that, when science deniers believe that the Earth is flat or that the MMR vaccine causes autism despite all of the evidence to the contrary, they are violating basic epistemic norms, and the violation of those norms would still be epistemically bad even if the resultant beliefs were entirely practically harmless.

Personally, I do not find this line of reasoning persuasive, as I doubt that epistemic norms have any independent normative force.Footnote ³⁹ However, since I cannot defend such a general epistemological view here, I will not pursue this objection. Instead, I will argue that, even if epistemic norms do have independent normative force, it does not follow that their violation is an appropriate object of concern on the part of the counter-deniers. Let me briefly mention two reasons why this is so.

First, just because my fellow citizens do something that I believe to be intrinsically bad, it does not follow that it would be appropriate for me to be concerned about it. Suppose, for example, that I believe that all lies are intrinsically morally bad. Does it follow that I should be concerned that many people lie? Clearly, it does not. After all, despite my beliefs about the morality of lying, I concede that many lies are perfectly harmless and that, in fact, some lies might even be beneficial (e.g., an innocuous lie told not to hurt a friend). The fact that something is intrinsically morally bad does not mean that it is all-things-considered bad, let alone an appropriate object of concern for others.

Second, all of us have some epistemically unjustified beliefs and most of us act (epistemically) irresponsibly from time to time. However, as long as our unjustified beliefs are practically harmless, our epistemic irresponsibility does not seem to be an apt object of concern on the part of others. For example, most people I know (myself included) seem to have relatively strong beliefs about the personality, the personal lives, or the moral character of various celebrities. These beliefs are usually based exclusively on what we have learned about those celebrities through the media. Most likely, most of those beliefs are epistemically unjustified (as they are based on little low-quality evidence) and, in holding them, we are likely acting in an (epistemically) irresponsible or blameworthy manner (since our beliefs should fit our evidence). However, no one seems particularly concerned about our epistemic failings in this regard, and rightly so. Even if most of our beliefs about celebrities are epistemically unjustified, our epistemic shortcomings do not seem grounds for concern if those beliefs are practically harmless.Footnote ⁴⁰ Similar considerations apply to more outlandish beliefs, such as the belief that people’s astrological signs affect their personalities. These beliefs seem no more epistemically justified than those of flat-Earthers, and those who hold them do not seem more epistemically responsible than flat-Earthers. However, no one seems to be particularly concerned about those beliefs, and rightly so (as long as the relevant beliefs are practically harmless). Why would the same considerations not apply to the beliefs of flat-Earthers?

As far as I can see, the most plausible answer to this question is that the beliefs of flat-Earthers and other science deniers, unlike beliefs about celebrities or beliefs about astrology, are not completely practically harmless – they are either directly or indirectly practically harmful. On closer inspection, thus, deontic epistemic concerns seem to be largely a distraction – the most serious concern about science denial is not that science deniers are epistemically incompetent or irresponsible but that their beliefs are directly or indirectly practically harmful. While it is still too early to assess the validity of the practical concerns, our preliminary survey suggests that it is plausible that some forms of science denial, such as vaccine hesitancy or HIV/AIDS denial, may be directly practically harmful and that it is not entirely implausible to suspect that forms of science denial that do not seem directly practically harmful (such as flat-Eartherism) may nonetheless be indirectly practically harmful, as they make those who embrace them more epistemically vulnerable to forms of science denial (or other unjustified beliefs) that are directly practically harmful.

2.1 Do We Need an Account of Science Denial?

This section explores one of the most basic questions about science denial – that is, what is science denial? Unfortunately, counter-deniers do not give a clear or univocal answer to this question. In fact, most counter-deniers do not even try to answer it at all. Instead, they simply rely on an open-ended list of prototypical forms of (alleged) science denial, which typically includes climate-change denial, vaccine hesitancy, creationism, or flat-Eartherism. But are such open-ended lists not sufficient? Why would we need an account of science denial? After all, we successfully use many ordinary terms (such as ‘chair’, ‘cat’, or ‘game’) without ever being given any explicit definition of them. We have all learned what a chair is by being exposed to a variety of chairs and being told that they were chairs. Why would the same not apply to ‘science denial’? Moreover, don’t we all know science denial when we see it?

There are several reasons to find this approach unsatisfactory. Let me mention four. First, it is not clear what is included in the open-ended list and what is not. While we typically agree on whether a given object should be classified as a chair, counter-deniers do not even agree among themselves on who should be classified as a science denier. Journalist Michael Specter, for one, interprets the concept so broadly as to include among science deniers not only anti-vaccination conspiracy theorists but also people who eat organic food or take multivitamins.Footnote ⁴¹ Presumably, most people (including most counter-deniers) would disagree with Specter on this point.Footnote ⁴² Whatever one may think of the health benefits of consuming organic foods or taking nutritional supplements, it seems implausible to tar the people who eat organic food or take multivitamins with the same brush as the most radical anti-vaxxers. However, without a clear notion of science denial, there is no principled way to determine where exactly to draw the line.

Second, without a clear account of science denial, it is unclear how prevalent science denial is or how concerned we should be about it. If, for example, we were to follow Specter in classifying anyone who eats organic food or takes nutritional supplements as a science denier, then science denial would be a very widespread phenomenon indeed. However, it would also not seem a particularly concerning one. After all, multivitamins and organic carrots might not have any beneficial effects on the well-being of those who consume them, but they do not seem directly or indirectly practically harmful either. The belief that organic carrots are healthier or more environmentally friendly than non-organic one appears to be practically and epistemically harmless even if false.

Third, even if counter-deniers could offer an exhaustive list of forms of science denial, this would only shift the problem – the problem would then be how to correctly use labels such as ‘climate-change denier’ or ‘anti-vaxxer’. For example, what qualifies someone as a climate-change denier? Must they deny that human activities have any impact on the climate, or can they merely deny that they have a significant impact? What about those who concede that human activities do have a significant impact on the climate but who minimize the seriousness of the effects of climate change? Or those who acknowledge the reality of anthropogenic climate change but claim that we should adapt to it rather than trying to mitigate it? The boundaries of climate-change denial don’t appear to be much clearer than those of science denial itself, so even if counter-deniers could offer an exhaustive list of all forms of science denial, it would still be unclear who counts as a science denier, as it would not be clear to whom each entry on the list applies.

Finally (and, to my mind, most importantly), unlike ‘chair’, ‘science denier’ is a term that has a strong evaluative and expressive content and, when such terms do not also have a clear descriptive content, they can be easily abused, and their abuse can itself be epistemically and practically harmful.Footnote ⁴³ Let me mention two examples of these harms. First, labels such as ‘science denier’ can be used to depoliticize political debates by dismissing our opponents’ opinions and concerns without properly engaging with them. For example, people concerned about the health risks of water fluoridation are often dismissed as science deniers and they are sometimes likened to General Jack Ripper, the paranoid character in Stanley Kubrick’s 1964 movie Dr Strangelove who believed water fluoridation to be a communist plot. However, a recent systematic review by the US National Toxicology Program suggests that concerns about water fluoridation may not be wholly unfounded.Footnote ⁴⁴ As its critics often point out, liberalism has a tendency to depoliticize political questions by portraying them as technical questions to be answered by experts.Footnote ⁴⁵ However, questions about water fluoridation policies are not merely technical, as they are partly questions about our preferences and values (such as the benefits and the risks of various water fluoridation policies). Moreover, branding certain concerns as science denial might have a chilling effect on scientific research, as researchers might fear that investigating them might negatively affect their professional reputation. The wholesale dismissal of concerns about water fluoridation as science denial, thus, is not only unwarranted given the evidence, but it is also risks stifling scientific inquiry, undermining policy debate, and delaying public policy action.Footnote ⁴⁶

Second, terms such as ‘science denial’ can be used to improperly politicize scientific questions.Footnote ⁴⁷ For instance, those labelled as science deniers often insinuate that the real science deniers are their opponents. In a 2015 interview, US Senator Ted Cruz, for example, claimed:

Cruz’s accusations mirror counter-deniers’ own charges and, absent a clear account of science denial, it is unclear how to prevent such abuses of the label.

These considerations suggest that, while we might not necessarily need a full-fledged definition of ‘science denial’ (or an account of science denial in terms of necessary and sufficient conditions), we still must gain a better understanding of the notion of science denial in order to apply it in a consistent and principled manner, to better understand the nature of the phenomenon it is meant to capture, or to assess the validity of the concerns it raises.

In this section, I argue that, once we do so, it becomes apparent that ‘science denial’ is a highly ambiguous term whose indiscriminate use conflates a heterogeneous set of distinct socio-epistemic phenomena. The more modest goal of this section is to clearly distinguish among some of these specific phenomena and to try to identify among them the one that is most likely to give raise to the counter-deniers’ concerns about science denial. Its more ambitious goal is to expose the inadequacy of a number of seemingly plausible ‘thin’ accounts of science denial – that is, accounts of science denial that describe the phenomenon without appealing to its explanation. In particular, this section argues that no ‘thin’ account of science denial can differentiate between science denial and several completely distinct phenomena while vindicating both the epistemic and the practical concerns of counter-deniers.

2.2 Science Non-Belief and Science Contestation

One of the most fundamental ambiguities of the term ‘science denial’ stems from two possible understandings of ‘denial’ in ordinary English. On the first understanding, ‘denial’ refers to a type of speech act – the act of denying that something or other is the case. For example, by uttering the words, ‘I have never met the victim’, the suspect performs the speech act of denying having ever met the victim. According to what we might call the expressive conception of science denial, science denial consists in the performance of certain speech acts – speech acts that deny the truth of the relevant scientific claims or, more generally, question their scientific standing (e.g., how well-supported they are by the evidence or how widely accepted they are in the relevant scientific community).Footnote ⁴⁸ To avoid confusion, in the rest of this Element, I refer to the performance of these kinds of speech acts in a given context as science contestation. So, for example, Donald Trump performed an act of science contestation by declaring on Twitter that ‘The concept of global warming was created by and for the Chinese in order to make U.S. manufacturing non-competitive’ (Donald J. Trump [@realDonaldTrump] Reference Trump2012).Footnote ⁴⁹ The expressive conception of science denial suggests a first possible account of science denial – one according to which a science denier is a science contester (i.e., someone who occasionally or systematically performs acts of science contestation in certain contexts).

On the second understanding, ‘denial’ refers to a psychological state – the state that is colloquially referred to as ‘being in denial’.Footnote ⁵⁰ Suppose, for example, that Andy consistently disregards the clear signs of his dependence on alcohol and the problems it causes him and those around him. As a result, Andy believes that he drinks only occasionally and that he could quit drinking anytime if he wanted to. Andy seems to be in a state of denial about his drinking habits. He does not believe that he has a drinking problem despite the fact that the evidence available to him supports the belief that he does.Footnote ⁵¹ According to what we might call the epistemic conception of science denial, science deniers are, like Andy, in denial about inconvenient truths.Footnote ⁵² They do not believe certain scientific claims despite the fact that the evidence available to them supports believing them. The epistemic conception of science denial suggests a second possible account of science denial – namely, one that identifies science denial with science non-belief, where, to a first approximation, a science non-believer is someone who does not believe certain scientific claims. On this account, someone who does not believe the relevant scientific claims qualifies as a science denier regardless of whether they disbelieve those claims (i.e., believe their negation) or whether they believe any related pseudo-scientific claims (such as the claim that the Earth is flat or that 5G cellphone towers cause COVID-19). However, those who actively disbelieve the relevant scientific claims or who believe pseudo-scientific claims automatically qualify as science non-believers, as those beliefs require not believing the corresponding scientific claims.

These two accounts of science denial disagree not only about the descriptive content of the term ‘science denier’ but also about its extension. While most science non-believers might also be science contesters, and vice versa, science non-belief and science contestation do not always go hand in hand. On the one hand, someone can be an insincere (science) contester – that is, a science contester who is not also a science non-believer. For instance, a politician who believes in anthropogenic climate change but claims not to believe in it for political gain would classify as an insincere contester. On the other hand, someone can be a tacit (science) non-believer – that is, a science non-believer who is not also a science contester. For example, someone who secretly harbours doubts about the safety of childhood vaccines but would never openly express them for fear of being branded an ‘anti-vaxxer’ would classify as a tacit non-believer.

Once science non-belief is distinguished from science contestation, it seems clear that, of these two notions, science non-belief is the more likely candidate for a disambiguation of ‘science denial’ as understood by counter-deniers. Unlike science contestation, science non-belief raises the most serious concern about science denial – namely, the concern that science denial might be directly practically harmful. The point can be seen most clearly by focusing on the ‘pure’ cases of each phenomenon – that is, tacit non-belief and insincere contestation. The practical concern applies directly to tacit non-belief but only indirectly to insincere contestation. The failure to believe in certain scientific claims might result in tacit science non-believers making decisions that have practically harmful consequences for themselves, for those around them, for their community, or even for the entire world. For example, to the extent that Ben’s failure to believe in the reality of anthropogenic climate change contributes to his decision to drive a highly fuel-inefficient vehicle or to oppose policies aimed at mitigating climate change, it seems to have practically harmful consequences for him and everyone else on the planet.Footnote ⁵³

Insincere contestation, on the other hand, does not seem directly practically harmful.Footnote ⁵⁴ Of course, this is not to deny that insincere contestation can be harmful. Science contestation is epistemically harmful to the extent that it pollutes the epistemic environment with misinformation that contributes to disseminating and sustaining science non-belief,Footnote ⁵⁵ and it is indirectly practically harmful to the extent that it contributes to disseminating or sustaining forms of science non-beliefs that are directly practically harmful. The literature on the anti-vaccine movement, for example, suggests that what I call science contestation pollutes the epistemic environment, thereby contributing to disseminating and sustaining unfounded and practically harmful beliefs about vaccines.Footnote ⁵⁶ Similarly, the literature on climate-change denial strongly suggests that what I have called science contestation (and, in particular, insincere contestation) contributes to disseminating and sustaining unfounded beliefs about climate change.Footnote ⁵⁷

However, the concerns about science contestation ultimately seem to boil down to practical concerns about science non-belief. More specifically, science contestation seems to be concerning only insofar as either it is a symptom of science non-belief (in its sincere form) or it contributes to disseminating and sustaining science non-belief (in both its sincere and insincere forms). Neither of these concerns, however, is about science contestation per se. Both are ultimately concerns about science non-belief. The arguments in Section 1 suggest that, if all forms of science non-belief were completely practically harmless, science contestation would not be an apt object of concern, even if it were a symptom of science non-belief or contributed to it. These considerations suggest that concerns about science denial are primarily concerns about science non-belief and only derivatively concerns about science contestation – science contestation is concerning either as a symptom of science non-belief or as a potential cause of it.

2.3 The Evidence Criterion and the Acceptance Criterion

On my preliminary definition, a science non-believer is someone who fails to believe specific scientific claims. This definition, however, leaves unanswered a crucial question – which scientific claims must one not believe to be classified as a science denier? Unless the class of relevant scientific claims is adequately restricted, the resultant account of science denial risks being overly inclusive, as all of us fail to believe some scientific claims, as we are all partake in what I have called ‘ordinary scientific ignorance’. So, which scientific claims must one fail to believe to qualify as a science denier?

Counter-deniers typically appeal to either of two general kinds of criteria to identify the relevant scientific claims. The first kind of criterion focuses on the first-order evidence for the scientific claims. One such criterion is that the relevant scientific claims are well corroborated (i.e., strongly supported by first-order empirical evidence).Footnote ⁵⁸ The second kind of criterion focuses on higher-order evidence for those claims. One such criterion is that the relevant scientific claims are widely accepted (i.e., accepted by a large proportion of the members of the relevant scientific communities).Footnote ⁵⁹ In what follows, I refer to the first criterion as the evidence criterion and to the second one as the acceptance criterion.Footnote ⁶⁰

To my knowledge, counter-deniers never discuss either criterion in detail or explain their reasons for adopting one criterion rather than the other. In fact, some counter-deniers seem to use the two criteria interchangeably.Footnote ⁶¹ This might be because they assume that the two criteria are always jointly satisfied – that is, that all well-corroborated scientific claims are widely accepted and vice versa. However, this assumption is false. For one thing, in some cases, scientific communities seem to accept claims based on insufficient evidence; and for another, scientific communities sometimes fail to accept a scientific claim despite its being well-corroborated. The case of the aetiology of peptic ulcers can be used to illustrate both points.Footnote ⁶² Until the late 1980s, the biomedical community widely accepted the hypothesis that peptic ulcers were caused by lifestyle factors (such as diet and stress). However, there was little evidence to support that level of scientific consensus.Footnote ⁶³ In the 1980s, two Australian physicians, Barry Marshall and Robin Warren, produced increasingly strong evidence that peptic ulcers were in fact caused by a bacterium called H. pylori.Footnote ⁶⁴ The biomedical community, however, initially ignored their findings. Increasingly frustrated, Marshall eventually decided to infect himself with H. pylori, developed peptic ulcers, and successfully treated them with antibiotics. It was only then that the scientific community started to take the bacterial hypothesis seriously and, eventually, accepted it. In 2005, Marshall and Warren were awarded the Nobel Prize in Medicine for their discovery.

As this case illustrates, the evidence criterion and the acceptance criterion can pull in different directions. For one thing, a scientific community can widely accept a scientific claim even when there is insufficient evidence for it (as the biomedical community did with respect to the lifestyle hypothesis) and, for another, it can fail to widely accept a scientific claim even when there is strong evidence for it (as it did with respect to the bacterial hypothesis). More importantly for us, this case shows that the evidence criterion and the acceptance criterion can disagree on who qualifies as a science denier. Someone who did not believe in the lifestyle hypothesis in the mid-1980s, for example, would count as a science denier according to the acceptance criterion but not according to the evidence criterion. So, which criterion should counter-deniers adopt?

The evidence criterion runs immediately into a number of serious problems. Here, I focus on two of them. The first is that it seems unable to distinguish between science denial and ordinary scientific ignorance. If a science denier is simply someone who does not believe some well-corroborated scientific claims, then we are all science deniers because each of us fails to believe a multitude of well-corroborated scientific claims. For example, presumably, most people do not believe that boron has two stable isotopes or that Arcturus is a red giant even if both claims are well corroborated. Most of us simply do not know that these claims are well corroborated. However, while we might all be scientifically ignorant to some extent, we are not all science deniers.Footnote ⁶⁵ For one thing, if we were all science deniers, then ‘science denier’ would be a completely useless label, as it would apply to everyone. For another, ordinary scientific ignorance seems epistemically blameless. Given the sheer size of our current scientific knowledge, nobody (not even the most knowledgeable scientist) can be expected to believe every well-corroborated scientific claim in every scientific field.

The second problem is that it seems highly unlikely that most non-experts would be capable of applying the evidence criterion on their own. Only the relevant experts seem capable of gathering and assessing the first-order evidence that bears on each scientific claim. For example, I would never be able to gather and assess first-order evidence that bears on the claim that boron has only two stable isotopes all on my own. For a non-expert like me, the best strategy to determine which scientific claims meet the evidence criterion would be to rely on the opinions of the relevant experts. However, this suggests that the only way for non-experts to comply as closely as possible with the evidence criterion is to employ a criterion such as the acceptance criterion – that is, to use the level of consensus among the experts on a given claim as higher-order evidence of the strength of the first-order evidence for that claim.

The acceptance criterion, however, also faces a number of serious problems. First, it is unclear how widely accepted a claim must be to meet the acceptance criterion. Does it have to be accepted by a simple majority of the community (50 per cent plus one) or by a supermajority (e.g., 95 per cent)?Footnote ⁶⁶ Second, it does not seem reasonable to expect non-experts to be able to apply the acceptance criterion either, as, for someone who is not familiar with the relevant scientific community, it is not easy to accurately assess how widely a given scientific claim is accepted in that community (or even identify the relevant scientific community).Footnote ⁶⁷ Third, the acceptance criterion too runs into the problem of distinguishing science denial from ordinary scientific ignorance. If a science denier is simply someone who does not believe a widely accepted scientific claim, then we are all science deniers because each of us fails to believe a multitude of widely accepted scientific claims (such as the claim about the stable isotopes of boron) out of epistemically blameless ignorance.

Fourth, the acceptance criterion would classify all scientific dissenters as science deniers, which seems particularly counterintuitive in the case of scientific mavericks whose unconventional scientific beliefs were eventually vindicated (such as Galileo, Darwin, or the aforementioned Marshall and Warren).Footnote ⁶⁸ In response to this problem, supporters of the acceptance criterion might argue that their account of science denial is only meant to apply to non-experts. This restricted account, however, seems to be inconsistent with the standard use of the label where it applies also to rogue experts such as Andrew Wakefield, the former physician who co-authored the infamous paper suggesting that the MMR vaccine might cause autism and who is now a prominent anti-vaccine activist. Another possible response is to adopt an account of science denial that applies the acceptance criterion to non-experts and the evidence criterion to experts. On this disjunctive account, a science denier is either an expert (in a relevant field) who does not believe some well-corroborated scientific claim or a non-expert who does not believe some widely accepted scientific claim. According to this account, a scientific dissenter qualifies a science denier only if they fail to believe well-corroborated scientific claims (within their area of expertise). This account might include scientific outcasts such as Wakefield among the science deniers while clearing scientific mavericks (such as Marshall and Warren) from the charge of science denial. However, it is not entirely clear that the disjunctive account is fully satisfactory. For one thing, it is unclear whether the beliefs of celebrated scientific mavericks are always well corroborated.Footnote ⁶⁹ For another, it is doubtful that experts can base their scientific beliefs exclusively on first-order evidence and not at all on the opinions of their scientific peers.Footnote ⁷⁰ Whatever the right account might be, these brief remarks suggest that the question of what criteria should be used to classify experts as science deniers may require a very different answer from the corresponding question about non-experts. Due to lack of space, I shall set the first question aside in the rest of this Element. In other words, the primary goal going is to identify an account of science denial that applies to non-experts but not necessarily to experts. Such an account might have to be combined with a separate account that applies to experts.

Fifth (and finally), it is dubious that any non-expert who fails to believe a widely accepted scientific claim should be classified as a science denier. Consider for example the case of Carla, a patient with peptic ulcers, who, in the mid-1980s, heard from her doctor about Marshall and Warren’s recently published research. Her doctor explained to her that, while the lifestyle hypothesis was widely accepted, two researchers had found credible evidence that a bacterium caused peptic ulcers and that they could be effectively treated with antibiotics. In light of this, Carla agreed to undergo a course of antibiotics, which successfully treated her ulcers. As a result, Carla came to disbelieve the lifestyle hypothesis while it was still widely accepted and believe the bacterial hypothesis instead. The acceptance criterion seems to entail that Carla was a science denier, that she remained one until the scientific community eventually caught up to her, and that, after that happened, she no longer was a science denier. In Carla’s case, however, these claims seem counterintuitive, as not only does her case not seem to be a case of epistemically blameworthy scientific ignorance – it does not seem to be a case of ignorance at all.

Hard-line supporters of the acceptance criterion might insist that this is wrong and that the epistemically responsible thing for non-experts is to believe all and only scientific claims that are widely accepted.Footnote ⁷¹ They might claim that, while Carla’s beliefs about peptic ulcers turned out to be true, it was epistemically irresponsible of Carla to disbelieve the lifestyle hypothesis while it was still widely accepted or to believe the bacterial hypothesis before it was widely accepted. The hard-line position, however, seems highly implausible. First, Carla’s beliefs did not have to be epistemically irresponsible. Suppose, for example, that Carla’s doctor had good reasons to believe that peptic ulcers were caused by H. pylori on the basis of Marshall and Warren’s research, that she explained those reasons to Carla as clearly and accurately as possible, and that Carla had good reasons to trust her doctor. Under those circumstances, Carla’s disbelief in the lifestyle hypothesis and her belief in the bacterial hypothesis would not seem epistemically irresponsible. In fact, it is usually assumed that non-expert should believe the testimony of trustworthy experts, and this is exactly what Carla did.Footnote ⁷²

Second, it seems unreasonable to insist that ordinary people should wait for a consensus to emerge to believe (or act upon) certain scientific claims. If, for example, there is significant evidence that gas stoves cause asthma in children,Footnote ⁷³ then this seems to be a sufficient reason for a parent to decide not to buy a gas stove even if there is still no scientific consensus over the claim that gas stoves cause asthma in children.Footnote ⁷⁴ In fact, such a parent can’t afford the luxury of waiting for a consensus to emerge before making their decision about whether they should replace their gas stove with an electric one.

Third, the hard-liners’ epistemic standards seem too demanding. If anyone who does not believe any widely accepted scientific claim is epistemically irresponsible, then we are all epistemically irresponsible because all of us fail to believe many widely accepted scientific claims out of ordinary scientific ignorance.

Moreover, the hard-line position seems to be even less tenable when applied to other cases. For instance, for much of the nineteenth and twentieth centuries, a broad scientific consensus held that upper-class white men were naturally endowed with greater intelligence than other groups, including women, racial and ethnic minorities, and lower-class individuals.Footnote ⁷⁵ According to hard-liners, those who rejected some of these widely accepted claims – including many first-wave feminists in the late nineteenth and early twentieth centuries – would qualify as science deniers. Yet, if science denial is epistemically blameworthy (or the result of epistemic irresponsibility), this seems to be a misdiagnosis. If anything, it was the scientists of that era, not the first-wave feminists, who were epistemically irresponsible and, arguably, blameworthy. In fact, such cases suggest not only that the failure to believe widely accepted scientific claims can be epistemically blameless (as with ordinary scientific ignorance) but that it may even be epistemically praiseworthy in some cases. Unlike the beliefs of the majority of scientists at the time, the beliefs of first-wave feminists were not only true but, arguably, also epistemically justified.Footnote ⁷⁶ If so, then first-wave feminists were justified (science) non-believers – they were epistemically justified in their disbelief of some widely accepted scientific claims.

Similar considerations apply to the case of lay experts (i.e., people who lack formal training or credentials in a given scientific field but who have developed substantial knowledge about it through lived experience, traditional culture, or practical engagement with its subject matter). Hard-liners would either have to deny the existence of lay experts (which is widely accepted among expertise scholars)Footnote ⁷⁷ or maintain that lay experts should always defer to credentialed experts (which is implausible, as it is not unusual for lay experts turn out to be right in cases of disagreement).Footnote ⁷⁸

The problems with the hard-line position might be avoided by modifying the acceptance criterion so as to exclude cases in which the scientific consensus is based on inadequate or spurious grounds (such as the racist, sexist, and classist prejudices of the relevant scientists). However, it is unclear that any general set of criteria can be used to identify all cases in which the scientific consensus is defective.Footnote ⁷⁹ Moreover, even if such criteria existed, it is dubious that ordinary people would be able to apply them effectively to real-world cases on their own.Footnote ⁸⁰ Finally, any such set of criteria could be misused to strategically disqualify genuine cases of scientific consensus in order to promote certain forms of science non-belief.

2.4 Science Disbelief

Counter-deniers might argue that the problem with the account of science denial as science non-belief is that it is too inclusive. Science denial should be identified with a more specific phenomenon – science disbelief. A science disbeliever is a non-expert who actively disbelieves some widely accepted scientific claim. Since all science disbelievers must be science non-believers but science non-believers need not be science disbelievers, this account of science denial is less inclusive and, as such, it might avoid some of the problems faced by the more inclusive account in terms of science non-belief. In particular, counter-deniers might hope that adopting this less inclusive account would allow them to distinguish science denial from ordinary scientific ignorance. After all, while most of us might not believe that boron has two stable isotopes, we do not disbelieve that claim either.

However, this account of science denial too faces some problems. Let me mention two. First, the account still fails to distinguish science denial from some of the other phenomena we have discussed. For one thing, many everyday beliefs are at odds with widely accepted scientific claims. Many people believe that cold weather causes colds, and most people believe that the question of whether any two events are simultaneous has an absolute, objective answer. However, while those beliefs are inconsistent with widely accepted scientific claims, they hardly warrant charges of science denial. For another, first-wave feminists disbelieved widely accepted scientific claims about the intellectual inferiority of women but, as argued earlier, should not be labelled science deniers.

Second, if the main concern of counter-deniers is that science denial is practically harmful, it is unclear why they should focus exclusively on science disbelief when science non-belief can be just as practically harmful (if not more). For example, someone who does not believe the claims that vaccines do not cause autism or that HIV causes AIDS may behave exactly as someone who actively disbelieves those claims. Those who produce scientific misinformation seem to know this all too well.Footnote ⁸¹ For example, the tobacco industry’s disinformation campaign concerning the health effects of second-hand smoking did not aim to persuade the public to disbelieve the relevant scientific claims (e.g., that second-hand smoking causes lung cancer). It had the more modest aim of trying to persuade them not to believe those claims. As a tobacco industry memo infamously put it, ‘doubt is our product’.Footnote ⁸² If the primary concern about science denial is its practical harmfulness, it is unclear why counter-deniers would focus exclusively on the narrower phenomenon of science disbelief rather than on the broader phenomenon of science non-belief, given that the latter is no less practically harmful than the former. In fact, some forms of science disbelief (such as flat-Eartherism) are arguably less practically harmful than some forms of science non-belief (such as certain forms of vaccine hesitancy).

2.5 Science Refusal

In response to these problems, counter-deniers may respond that science denial is not merely a failure to believe widely accepted scientific claims – it is a refusal to believe them. While all of us fail to believe many such claims, most of us do not actively refuse to believe them. On this account, a science denier is a science refuser – that is, a non-expert who refuses to believe some widely accepted scientific claim.Footnote ⁸³ Yet, it remains unclear what it means to refuse to believe a claim (as opposed to merely failing to believe it).

Earlier, I suggested that the reason why ordinary scientific ignorance may be epistemically blameless is that most of us simply do not know that the relevant scientific claims are widely accepted. This inspires the following proposal. A science refuser is a non-expert who does not believe some widely accepted scientific claim despite knowing that it is widely accepted. However, this proposal encounters at least two difficulties. First, someone might know that a scientific claim is widely accepted and yet have good reasons to suspect that the scientific consensus is defective – for example, unreliable research,Footnote ⁸⁴ or community biases,Footnote ⁸⁵ or other systemic failures.Footnote ⁸⁶ Consider again the case of first-wave feminists who rejected the scientific consensus on the intellectual inferiority of women. While they did refuse to believe claims that, at the time, were widely accepted, labelling them as science deniers seems wrong if the label implies that they were epistemically irresponsible or blameworthy. Second, this proposal excludes many of those who counter-deniers would pre-theoretically classify as science deniers. Empirical evidence suggests that many alleged science deniers do not believe that the relevant scientific claims are widely accepted,Footnote ⁸⁷ which, on this proposal, entails that they are not science deniers after all. This proposal is thus both overly inclusive and overly narrow.

Counter-deniers might respond that, even if some science deniers might not know that the relevant scientific claims are widely accepted, they should know. However, this proposal is still vulnerable to the first objection: first-wave feminists knew that a scientific consensus existed regarding the intellectual inferiority of women, but (correctly) doubted the grounds of that consensus. Moreover, as argued earlier (Section 2.3), it is unrealistic to expect non-experts to be able to know how widely accepted a scientific claim is. Finally, it is unclear whether this proposal can distinguish science denial from ordinary scientific ignorance. After all, why should non-experts know whether there is a scientific consensus on the causes of climate change but not whether there is one on, say, the causes of the extinction of dinosaurs?

Counter-deniers might reply that, in the case of anthropogenic climate change, the duty is not a purely epistemic one but (partly) a moral or political one. People should know that there is a scientific consensus on the causes of climate change because those scientific claims are relevant to their practical decisions as citizens or as consumers while the causes of the extinction of dinosaurs have no practical relevance for them. However, this answer too faces problems. First, even if there was such a duty to know, it would not seem to apply to scientific claims that are not (directly) relevant to people’s practical decisions, such as the claim that Earth is a globe or the claim that it is more than 10,000 years old. Second, the first-wave feminist case suggests that there is no moral or political duty to accept the scientific consensus on morally or politically charged issues. Finally, while many people might know that there is a broad scientific consensus about the reality of anthropogenic climate change, very few people seem to know whether or not there is an equally broad scientific consensus over other crucial practically relevant questions about anthropogenic climate change (such as the existence of so-called ‘tipping points’ or the likely consequences of climate change under various possible scenarios). If the supposed duty to know is triggered by the practical relevance of that knowledge, then wouldn’t these be things all of us have a duty to know? And, if we don’t, are we all science refusers?

While the suggestion that science deniers refuse (rather than merely fail) to believe widely accepted scientific claims might seem promising, without a clear account of science refusal, it merely replaces the obscure notion of science denial with the no-less-obscure notion of science refusal.

2.6 Epistemically Blameworthy Science Non-Belief

Counter-deniers might hope to avoid all of the problems discussed so far simply by identifying science denial with (epistemically) blameworthy (science) non-belief. On this account, a science denier is a non-expert whose failure to believe some widely accepted scientific claim is epistemically blameworthy. At first, the blameworthy non-belief account might seem promising. For one thing, it would seem to solve the problem of distinguishing science denial from epistemically blameless forms of science non-belief (such as ordinary scientific ignorance or justified science non-belief). However, on closer scrutiny, its promises appear illusory.

First, unless the account is complemented by an explicit account of epistemic blameworthiness, it is of little use. For one thing, it provides no independent criterion to determine who counts as a science denier and, without an independent criterion, any classification would seem somewhat circular. Would counter-deniers classify Ben as a science denier only after having determined that his beliefs about climate change are epistemically blameworthy or would they simply assume that his beliefs about climate change are epistemically blameworthy because he does not believe in the reality of anthropogenic climate change? For another, each of us would have to rely on their own intuitions to determine whether a given case of science non-belief is epistemically blameworthy, intuitions that may be neither fully dependable nor universally shared.

Second, ideally, an account of science denial should not simply state that science denial is epistemically blameworthy – it should also explain why it is epistemically blameworthy. However, this account not only fails to do so – it does not even try. The account seems to suggest that the alleged epistemic blameworthiness of science denial is a mystery that cannot be further explained or investigated.

Finally, some cases of epistemically blameworthy non-belief do not seem to qualify as science denial. Consider, for example, what we might call indifferent (science) non-belief. Had I studied chemistry more diligently in high school (or paid better attention in class), I would now believe many widely accepted scientific claims that I do not currently believe. Arguably, my ignorance of these claims is epistemically blameworthy. I would seem worthy of epistemic blame for failing to believe widely accepted scientific claims that I could have easily believed had I taken my epistemic responsibilities more seriously. Yet, while indifferent non-belief seems to be epistemically blameworthy, it clearly does not deserve to be classified as science denial.

2.7 Science Non-Compliance

In light of the problems faced by epistemic accounts of science denial, counter-deniers might consider adopting an account that focuses on the practical behaviour of science deniers rather than on their epistemic behaviour. In particular, counter-deniers might adopt an account according to which a science denier is a science non-complier – that is, a non-expert who does not not to follow recommendations or regulations based on widely accepted scientific claims. This approach appears to avoid some of the problems plaguing epistemic accounts. Most importantly, it seems to distinguish science denial from ordinary scientific ignorance. While all of us fail to believe many widely accepted scientific claims, most of us are still willing to follow recommendations or regulations based on those claims, which means that we would not qualify as science deniers under this account.

However, the account is obviously overly inclusive. Science non-compliance can be the result of a multiplicity of factors that seem to have little or nothing to do with science denial. For example, in the United States, uninsured patients have significantly lower vaccination rates than insured ones.Footnote ⁸⁸ This suggests that limited access to healthcare and other socio-economic factors play a significant role in many cases of science non-compliance. In other cases, science non-compliance might be due to prudential considerations. Consider, for example, the case of Daphne, who disbelieves allegations that the MMR vaccine causes autism and who is aware that public health guidelines recommend that children her sons’ age should receive their first dose of the MMR vaccine. However, her belief that the MMR vaccine does not cause autism is not strong enough to overcome her fear that her son might develop autism.

For the sake of clarity, it is convenient to frame this phenomenon in decision-theoretic terms. Suppose that Daphne attaches a much greater disutility to the very unlikely (even by her own lights) possibility that her son might develop autism due to the MMR vaccine than to the much more likely possibility that he might contract measles, mumps, or rubella as a result of not being vaccinated against those diseases. If the discrepancy between her utilities is large enough, then the expected utility of not vaccinating her child will exceed the expected utility of vaccinating him even if her credence in in the claim that the MMR vaccine causes autism is much lower than her credence in its negation. In fact, according to expected utility theory, given Daphne’s beliefs, and her preferences, it would be rational for her to decide not to vaccinate her child.Footnote ⁸⁹

The key point is that, while Daphne is a science non-complier, she is not a science non-believer, as she does believe that the MMR vaccine does not cause autism. Rather, Daphne is what we might call a prudential (science) non-complier. In her case, prudential considerations trump epistemic ones. It should be also noted that, while, in some cases, the preferences of prudential non-compliers might be the result of prejudices (such as ableist prejudices about autism) and other inaccurate factual beliefs (such as beliefs about the seriousness of diseases such as measles, mumps, or rubella), this need not always be the case. In some cases, it might simply be based on a simple discrepancy between the values of the prudential non-compliers and those on which the relevant recommendations or regulations are based.

Crucially, while attempts to correct the beliefs that underpin Daphne’s preferences might be effective in some cases of prudential non-compliance, attempts to increase her credence in the claim that the MMR vaccine causes autism might have no effect on her behaviour as long as prudential considerations trump epistemic ones. In the most extreme cases, prudential non-compliance becomes what we might call prudential swamping – the disparity in the utilities of the outcomes is so large that the expected utility of non-compliance exceeds that of compliance no matter how high the subject’s credence in the relevant scientific claim is (as long as is not one). Again, the only effective approach in cases of prudential swamping might be to address any mistaken factual beliefs underlying the prudential swamper’s preferences.

It is worth noting that prudential non-compliance is not a purely theoretical possibility. Some widely discussed empirical studies seem to show that convincing vaccine hesitators of the safety and effectiveness of vaccines does not necessarily increase their intent to be vaccinated (or have their children vaccinated).Footnote ⁹⁰ This sort of result is often interpreted as evidence of the stubbornness of science deniers.Footnote ⁹¹ However, the notion of prudential non-compliance offers an alternative explanation that does not postulate an epistemic failing on the part of these vaccine refusers – namely, the decisive factor in their decision to refuse vaccines is not their science non-belief, but their prudential-science non-compliance.

Counter-deniers might argue that these problems can be avoided by restricting the account so as to include only non-experts who, unlike Daphne, are both science non-believers and science non-compliers. However, this account still faces a number of problems. First, according to it, prototypical examples of science denial such as flat-Eartherism or creationism do not qualify as cases of science denial, as neither system of beliefs requires those who adopt them not to follow any recommendations or regulations based on widely accepted scientific claims.

Second, it is not easy for a non-expert to determine whether a piece of scientific advice is based on scientific claims that are widely accepted. For example, early in the COVID pandemic, the US Centers for Disease Control advised people not to wear face masks unless they were ill. However, for most non-experts at the time, it would have been unclear whether that piece of advice was based on widely accepted scientific claims about the effectiveness of face masks against COVID-19 (rather than on concerns about equipment shortages for healthcare workers). If, as I argued earlier, it is unreasonable to expect non-experts to be able to figure out whether a given scientific claim is widely accepted, it seems even more unreasonable to expect them to be able to figure out which set of scientific claims each piece of scientific advice they receive is based on and whether each of those claims is widely accepted.

Third, there are many cases in which recommendations are based on scientific claims that are widely accepted but false. For example, in the 1980s, the recommendation for Carla would have been not to use antibiotics to treat her peptic ulcers. However, as I argued earlier, it would seem wrong to classify Carla as a science denier even if she did not follow recommendations that were based on scientific claims that were widely accepted at the time.

Finally, no recommendation or regulation is ever based exclusively on scientific claims. Recommendations or regulations always also presuppose value judgments. So, it is possible that a science non-believer who does not follow a piece of scientific advice would still not follow it even if she were to come to believe the widely accepted scientific claims on which it is based. This seems to suggest that, even in those cases in which science non-compliance is partly the result of science non-belief, there is no guarantee that addressing science non-belief would lead to the former science non-believers making less practically harmful decisions, as they might also be prudential non-compliers. As I argue in more detail in the next section, part of the challenge in assessing the validity of the counter-deniers’ practical concerns is that it is often extremely difficult to determine whether or not science non-belief plays any role in someone’s making practically harmful decisions and, even more so, whether or not it plays a decisive role – that is, whether the science non-believers would make different practical decisions if it were not for their science non-belief.

2.8 Conclusion

This section explored several ‘thin’ accounts of science denial and found them all inadequate. While, admittedly, these arguments fall short of demonstrating that no ‘thin’ account of science denial is tenable, they do seem to show that the task of formulating a ‘thin’ account of science denial is much harder than counter-deniers seem to think. A much more promising option is to turn instead to ‘thick’ accounts of science denial – that is, accounts of science denial that describe the phenomenon by explaining its causes. The next section turns to the most promising and prominent account of this kind, which explains science denial in terms of motivated reasoning.

3.1 Motivated Science Non-Belief

The previous section argued that ‘thin’ accounts of science denial (i.e., accounts of science denial that attempt to describe the phenomenon without explaining it) face very serious problems. This section turns to what is arguably the most promising and prominent ‘thick’ account of science denial, which claims that science denial results from motivated reasoning.Footnote ⁹²

Motivated reasoning is ‘the tendency of individuals to unconsciously conform assessment of factual information to some goal collateral to assessing its truth’ (Kahan Reference Kahan, Robert and Kosslyn2016, 2). More precisely, someone engages in motivated reasoning about a given issue whenever the aim of their inquiry into that issue is to achieve some non-epistemic goals of theirs at the expense of epistemic goals, such as the goals of acquiring true beliefs or avoiding false ones. For the sake of clarity, it is helpful to distinguish distal goals from proximate goals, where a (more) proximate goal is an intermediate goal that is instrumental to the achievement (more) distal goals and one’s ultimate goals are one’s most distal goals. When I double-check that my appointment with the dentist is at 3 p.m., my distal goals might be exclusively non-epistemic (e.g., the goal of not being late for my appointment). However, if my inquiry into the claim that my appointment is at 3 p.m. is to contribute to achieving my distal non-epistemic goals, it must also aim to achieve the proximate epistemic goal of having a true belief about the time of my appointment. So, even if all of my distal goals are non-epistemic, I am not engaging in motivated reasoning, as my inquiry does not aim to achieve my distal non-epistemic goals at the expense of the proximate epistemic goal of having a true belief about the time of my appointment. In fact, the opposite is true, as I realize that it is only by achieving the proximate epistemic goal of having a true belief about the time of my appointment that I can achieve my distal non-epistemic goal of not missing it. If, on the other hand, my only goal in inquiring about the time of my appointment is the non-epistemic goal of placating my anxiety about being late to my appointment (regardless of whether my appointment is, in fact, at 3 p.m.), then I would be engaging in motivated reasoning, as my inquiry about that claim would aim to achieve a non-epistemic goal at the expense of the epistemic goal of having a true belief about the time of my appointment. Similarly, Andy engages in motivated reasoning about his drinking habits if his inquiry into the topic aims to achieve non-epistemic goals (e.g., the goal of feeling better about himself) at the expense of epistemic goals, such as that of avoiding false beliefs about his drinking habits.

The motivational account of science denial maintains that a science denier is a motivated (science) non-believer – that is, someone who fails to believe some widely accepted scientific claim as a result of motivated reasoning.Footnote ⁹³ On this account, people who do not believe in the reality of anthropogenic climate change or the safety of childhood vaccines fail to do so because their reasoning about those issue aims to achieve some non-epistemic goals of theirs (e.g., the goals of protecting their political ideology or that of affirming their social identity) at the expense of any epistemic goals (e.g., the goal of having true beliefs about those issues). As philosopher Lee McIntyre puts it:

The motivational account has several features that counter-deniers are likely to find appealing. Let me mention six here. First, if motivated reasoning is inherently epistemically irresponsible (or blameworthy),Footnote ⁹⁵ then the motivational account can explain why science denial is also epistemically irresponsible (or blameworthy), thus validating the counter-deniers’ deontic epistemic concerns.Footnote ⁹⁶ Second, the motivational account can easily distinguish science denial from other forms of science non-belief, such as ordinary scientific ignorance, justified non-belief, and indifferent non-belief. Unlike those other forms of science non-belief, science denial is the result of motivated reasoning. Third, the motivated account seems to apply to experts as much as non-expert. A scientific expert who refuses to believe widely accepted scientific claims due to motivated reasoning is as much a science denier as a non-expert that does the same. Fourth, if, as the Andy example shows, the state of being in denial can be construed as a product of motivated reasoning, the motivational account can also explain in what sense science deniers are in denial. Fifth, the motivational account can explain in what sense science deniers can be regarded as being wilfully ignorant about the relevant scientific claims or as refusing to believe those claims (as opposed to merely failing to believe them). It seems natural to describe someone who does not believe a claim as a result of motivated reasoning in those ways. For example, if Andy does not believe that he has a drinking problem because his reasoning about the matter aims to achieve the non-epistemic goal of not feeling bad about himself at the expense of the epistemic goal of having true beliefs on the matter, then it would be natural to describe him as refusing to believe that he has a drinking problem or as being wilfully ignorant of his drinking problem. Finally, the motivational account supports the post-truth diagnosis of the collective epistemic malaise afflicting liberal democracies. As mentioned earlier (Section 1.3), the post-truth diagnosis attributes that malaise to a lack of interest in (or concern for) the truth (and other epistemic goals) on the part of a growing number of our fellow citizens. If one of the main symptoms of that malaise (i.e., science denial) turns out to be the result of reasoning that aims to achieve non-epistemic goals at the expense of epistemic ones, then this would provide strong support for the post-truth diagnosis. The motivational account of science denial can thus be seen as a crucial test case for the post-truth diagnosis.

It is important to note, however, that, in order to have all of the aforementioned features, the motivational account must be combined with an account of motivated reasoning that differs significantly from the one typically adopted by cognitive scientists.Footnote ⁹⁷ Cognitive scientists who work on motivated reasoning are primarily interested in studying how non-epistemic factors such as values, social identities, and practical goals affect human cognition. What I shall call the universalist account of motivated reasoning views motivated reasoning as a universal, involuntary, and unconscious feature of human cognition. On the universalist account, the beliefs of those who believe widely accepted scientific claims about the reality of anthropogenic climate change or the safety of childhood vaccines are just as likely to be unconsciously influenced by their values, their social identities, or their practical goals as the beliefs of those who do not believe in those scientific claims. The universalist account is both normatively symmetric and normatively neutral. It is normatively symmetric because those who believe in the reality of anthropogenic climate change or the safety of childhood vaccines rely on motivated reasoning just as much as those who do not, so the ones are no more epistemically irresponsible or blameworthy than the others, and it is normatively neutral because, since motivated reasoning is both unconscious and involuntary feature of human cognition, it would be inappropriate to hold either side responsible for engaging in it or blame them for it.

In order to have the aforementioned features, the motivational account must thus be combined with a particularist account of motivated reasoning instead of a universalist one. On the particularist account, motivated reasoning is neither universal nor entirely involuntary or unconscious. It is a kind of reasoning people employ in a not-entirely-involuntary and not-entirely-unconscious endeavour to preserve their own ignorance on certain topics. The particularist conception of motivated reasoning is thus much closer to the notion of wilful ignorance employed by philosophers than to the universalist notion of motivated reasoning employed by cognitive scientists. The particularist account holds that people engage in motivated reasoning in an attempt to remain ignorant of truths that are inconvenient to them, which is why their ignorance is wilful. It is the fact that their use of motivated reasoning is neither entirely involuntary nor entirely unconscious that makes the resulting ignorance both wilful and epistemically culpable. On this account, motivated science non-believers employ motivated reasoning to wilfully ignore inconvenient truths such as the reality of climate change, which makes their epistemic conduct epistemically irresponsible (and blameworthy). As these remarks suggest, the motivational account has all the features mentioned earlier only when it is combined with a particularist account of motivated reasoning. It is for this reason that, unless otherwise stated, in what follows, I use ‘motivated reasoning’ in its particularist sense, not the universalist one. While many of the arguments in this section apply also to those who adopt a universalist account, the main target are counter-deniers, who adopt a particularist one.

This section argues that, although the account of science denial as motivated non-belief fares much better than ‘thin’ accounts of science denial discussed in the previous section, it does not vindicate all of the concerns of counter-deniers. Once we carefully distinguish motivated non-belief from other superficially similar phenomena, it seems dubious that any of them warrants all the concerns that animate counter-deniers. More specifically, I argue that counter-deniers face a dilemma. On the one hand, even if motivated non-belief might be epistemically irresponsible (or blameworthy), it is doubtful that it is ever practically harmful. On the other hand, even if other forms of science non-belief may be sometimes practically harmful, it is doubtful that they are always (or ever) epistemically irresponsible (or blameworthy).

3.2 Misinformed Science Non-Belief

This subsection argues that there is a significant tension between motivated reasoning and another factor often implicated in the aetiology of science denial – exposure to scientific misinformation. Even counter-deniers typically concede that exposure to scientific misinformation can contribute to the science deniers’ failure to believe certain scientific claims.Footnote ⁹⁸ In fact, as we saw in the previous section, one of the main concerns about science contestation is that it can have epistemically harmful consequences for those who are exposed to it – it can turn them into science non-believers.

For our purposes, it is useful to distinguish motivated non-believers (who do not believe widely accepted scientific claims due to motivated reasoning) from misinformed (science) non-believers (who do not believe widely accepted scientific claims exclusively due to exposure to misinformation). A few remarks are in order here.

First, this distinction does not deny that misinformation plays a role in motivated non-belief as well. However, misinformation plays distinct roles in motivated non-belief and misinformed non-belief. In particular, exposure to misinformation (as understood here) is a passive phenomenon, which must be distinguished from the active pursuit of misinformation often associated with motivated reasoning. If Evan engages in motivated reasoning about climate change to protect his political beliefs, he might actively seek out (whether consciously or not) misinformation that supports his beliefs about climate change. In my terminology, Evan is not exposed to misinformation, as his willingness to trust the sources of that misinformation partly depends on the content of the misinformation they provide to him – he is willing to trust a given source in part because it provides him with misinformation that supports his beliefs about climate change. When Evan seeks out misinformation, his encounters with misinformation are a result of motivated reasoning, not accidental exposure. Frank, on the other hand, has been exposed to misinformation about climate change from a variety of sources that he independently trusts, such as his favourite news channel, his favourite radio talk-shows, his favourite politicians, his pastor, his friends, and his family. Crucially, his willingness to trust those sources, unlike Evan’s, is (largely) independent of the content of the misinformation they spread. He trusts those sources on independent grounds, not because they spread the kind of misinformation he was looking for as in the case of Evan. For example, Frank might trust his friends because he has independent first-hand evidence of their trustworthiness in other domains, and he trusts his favourite news channel because it enjoys a good reputation among those he trusts. In my terminology, while Frank is exposed to misinformation, Evan is not – he (actively) seeks out misinformation rather than being (passively) exposed to it.

Second, the fact that Evan is a motivated non-believer (about climate change) does not mean that he cannot be also a misinformed non-believer (on the same topic). In fact, given his political beliefs and the social network in which he is embedded, it is likely that Evan already trusts sources that spread misinformation about climate change independently of the fact that they spread misinformation about it. If this is so, then Evan is an ‘impure’ motivated non-believer – his science non-belief is due to a combination of motivated reasoning and exposure to misinformation. While it is likely that most actual motivated non-believers are ‘impure’ motivated non-believers,Footnote ⁹⁹ for the sake of clarity, I focus primarily on ‘pure’ cases here – cases in which motivated non-believers are not at all exposed to misinformation (even if they might actively seek it out and find it).

Third, even if there are strong reasons to suspect that most motivated non-believers in the real world are ‘impure’ motivated non-believers, there seems to be no reason to think that there are no misinformed non-believers out there.Footnote ¹⁰⁰ In fact, it seems highly plausible to assume that some people fail to believe widely accepted scientific claims exclusively as a result of their exposure to misinformation without motivated reasoning playing any role in their science non-belief. Consider, for example, a trivial case of misinformed non-belief – that of Geeta, a high school student who, due a typo in her chemistry textbook, comes to mistakenly believe that uranium-236 (not uranium-238) is the most abundant naturally occurring isotope of uranium. Clearly, Geeta is a misinformed non-believer, as her false belief about the isotopes of uranium does not serve any non-epistemic goals of hers.Footnote ¹⁰¹ However, if Geeta is a misinformed non-believer, then it is not clear why Frank would not also be one. After all, his case differs from hers only in degree, not in kind. While Frank has been exposed to much more misinformation from a much wider variety of trusted sources than Geeta, both are misinformed non-believers.

Once misinformed non-belief is carefully distinguished from motivated non-belief, it should be obvious that misinformed non-believers should not be classified as science deniers. People who do not believe a given claim as a result of their exposure to misinformation are typically not in denial about that claim. For example, if you mistakenly believe that your flight is boarding because the airport monitors erroneously said so, you are not in denial about the fact that your flight is not boarding yet – you have been simply misinformed about the status of your flight. Moreover, many forms of misinformed non-belief seem completely epistemically blameless. Geeta, for example, is a misinformed non-believer, but her failure to believe that uranium-238 (not uranium-236) is the most abundant naturally occurring isotope of uranium does not make her a science denier, as she has no motive to refuse to believe that particular claim or be in denial about it.

As this last example shows, questions about epistemic responsibility bring into sharper focus the tension between these two diagnoses. While motivated non-belief might be always epistemically irresponsible (or blameworthy), it is doubtful that misinformed non-belief is typically epistemically irresponsible (or blameworthy). To see why, consider the analogous case of someone who falls for a scam. While the moral responsibility for scams might always fall squarely on the scammer, apportioning the epistemic blame is a bit more complicated. In the case of someone who falls for a blatant scam (such as a ‘Nigerian prince’ scam), it might be tempting to attribute some of the epistemic responsibility to the victim, as they could be blamed for being gullible and gullibility is a form of epistemic irresponsibility. However, in the case of a sophisticated scam (such as a scam pulled by a group of con artists who have carefully cultivated their victim’s trust over a long period of time), the epistemic responsibility seems to fall exclusively on the scammers, as we recognize that, in such a case, it would be unreasonable to expect the victim to be able to realize that they are falling victim to a scam. From an epistemological point of view, many cases of misinformed non-belief seem much closer to the sophisticated scam than to the blatant one. Frank’s case, for example, seems to be much closer to the case of the victim of a sophisticated scam by a group of con artists who have cultivated his trust over an extended period of time than to that of the victim of a blatant scam. If so, then Frank is neither epistemically irresponsible nor worthy of epistemic blame.

Moreover, if Frank is epistemically irresponsible (or blameworthy), it is because he placed his trust in sources that are not trustworthy (at least with regard to climate change). However, while, sometimes, it might be epistemically irresponsible or blameworthy to place one’s trust in untrustworthy sources (e.g., when one has reason to doubt their trustworthiness), this is not always the case. Our epistemic environment is increasingly flooded with both information and misinformation and it is practically impossible for any of us to discern one from the other all on our own. When it comes to information such as scientific information that we have no way to independently verify on our own, we seem to have no other realistic option but to trust the sources that have a good reputation within our own community and mistrust those that do not (just like Frank does).Footnote ¹⁰² This makes us epistemically dependent on our own communities. If we are fortunate enough to inhabit an epistemically discerning community (i.e., one in which the reputation of various sources of information reliably tracks their actual trustworthiness on each of the topics they cover), then we tend to trust trustworthy sources and mistrust untrustworthy ones.Footnote ¹⁰³ If, on the other hand, we have the misfortune to inhabit a community that is not epistemically discerning, then we might trust untrustworthy sources and mistrust trustworthy ones. The crucial difference between the good case and the bad one seems to be a difference in epistemic luck, not epistemic responsibility. It would seem unreasonable to expect Frank to mistrust sources that are well reputed in his community and trust ones that are ill-reputed unless he has independent reason to doubt that his community is epistemically discerning, and it is at best unclear that Frank could ever have any such reason.Footnote ¹⁰⁴

Finally, even if, contrary to the previous argument, it was epistemically irresponsible for Frank to place his trust in those sources, the epistemic irresponsibility involved would be different in both degree and kind from the epistemic irresponsibility that is allegedly involved in motivated non-belief. At most, a misinformed non-believer like Frank has chosen the wrong means to achieve his epistemic goals; a motivated non-believer like Evan, on the other hand, is not trying to achieve any epistemic goals in the first place.Footnote ¹⁰⁵

These arguments show that, while, in the real world, it might not always be easy to discern misinformed non-belief from motivated non-belief, we must carefully distinguish the one from the other if we want to better understand (or address) either phenomenon, as there are crucial differences between them. On closer scrutiny, cases of misinformed non-belief seem much closer to cases of ordinary scientific ignorance than to cases of science denial (whether understood as motivated non-belief or in the pre-theoretical sense).

3.3 Misinformed Non-Belief in High-Stakes Cases

Once misinformed non-belief is carefully distinguished from motivated non-belief, it is easier to see why the counter-deniers’ practical concerns may be largely misplaced. It is because misinformed non-belief is a much more plausible diagnosis than motivated non-belief in high-stakes cases of alleged science denial – that is, cases in which science non-belief contributes to decisions that have serious practically harmful consequences for the science non-believer or their loved ones.

Consider, for example, the case of Helen, an apprehensive first-time parent who has concerns about vaccinating her daughter. After her concerns are hastily dismissed by her overworked family doctor, Helen researches the topic online and stumbles upon a support group for parents who believe their children have developed autism as a result of vaccines. The parents in the group, who, unlike her doctor, seem to genuinely care about the well-being of Helen’s daughter, tell Helen about how their children were diagnosed with autism soon after receiving their vaccinations and point her to a number of online resources that claim that childhood vaccines can cause autism and several other serious conditions. As a result, Helen comes to believe that childhood vaccines cause autism and decides not to have her daughter vaccinated.Footnote ¹⁰⁶

It is highly implausible to suggest that Helen is a motivated non-believer. In fact, we can plausibly assume that Helen’s ultimate goal in deciding whether or not to believe that vaccines cause autism is to keep her daughter safe and healthy; and, while, admittedly, that’s not an epistemic goal, we can also plausibly assume that Helen realizes that, in order to pursue it, she must pursue proximate epistemic goals such as having true beliefs about the safety of vaccines that are instrumental to its achievement. This seems to suggest that, if Helen’s conduct is epistemically irresponsible, it is not because her reasoning about the safety of vaccines aims to achieve her non-epistemic goals at the expense of epistemic ones. Not only her distal non-epistemic goals are compatible with her epistemic goals, but Helen realizes that her epistemic goals are instrumental to her non-epistemic goals. Nor is it because Helen actively sought out misinformation about the topic, as Helen stumbled into the online support group accidentally and trusted them more than her doctor because, unlike him, they genuinely seemed to care about the well-being of her child. If Helen’s conduct is epistemically irresponsible, it is because she has employed epistemic means that are inadequate to the achievement of her epistemic goals. She tried to assess the truth of the claim that vaccines cause autism by relying on sources that are not trustworthy on that topic.Footnote ¹⁰⁷ But this means that Helen is a misinformed non-believer, not a motivated non-believer.

Moreover, it is not even clear what non-epistemic goals Helen would achieve by believing that vaccines cause autism. If motivated reasoning is primarily used to protect one’s wilful ignorance of inconvenient truths (as in Andy’s case), then it is implausible to suggest that Helen is engaging in motivated reasoning about the safety of vaccines or that she is in denial about it. The suggestion that Helen is in denial would make much more sense if there was no safe and effective way to protect her daughter from a serious illness but, nevertheless, Helen believed that she could do so by performing a religious ritual or giving her homeopathic remedies. However, it is unclear why Helen would be in denial about the existence of a safe and effective way to protect her daughter from several serious illnesses, as, if anything, that is the opposite of an inconvenient truth.

While not all vaccine-hesitant parents might have the best interest of their children at heart, presumably, most of them do. If so, then it is plausible to assume that most parents who do not believe that childhood vaccines are safe are misinformed non-believers like Helen rather than motivated non-believers. If the ultimate goal of most vaccine-hesitant parents is keeping their children safe and healthy, they presumably realize that, in order to achieve that goal, they must pursue epistemic goals that are instrumental to it, such as the goals of acquiring true beliefs about the safety of vaccines and avoiding false ones. In turn, this means that it is highly implausible to assume that they engage in motivated reasoning about the safety of vaccines. It is much more plausible to assume that, like Helen, those parents are misinformed non-believers – their mistaken beliefs about vaccines are the result of employing ineffective epistemic means to achieve their epistemic goals (e.g., trusting unreliable sources of information about vaccines and mistrusting trustworthy ones), not of pursuing of non-epistemic goals at the expense of epistemic ones.

This argument seems to generalize to other prototypical high-stakes cases of (alleged) science denial, such HIV/AIDS denial, or certain forms of COVID scepticism. In those cases, too, it seems highly implausible to suggest that the alleged science deniers do not believe the relevant scientific claims due to motivated reasoning. Presumably, in high-stakes cases, people are trying their utmost to make decisions that are in their best interest (or that of their loved ones). Consider the case of Ivan, an HIV-positive patient who refuses antiretroviral drugs because he believes AIDS was invented by pharmaceutical companies to sell those drugs. Presumably, Ivan is not in denial about the truth of the claim that antiretroviral drugs are a safe and effective way to prevent or reverse AIDS, as that is not an inconvenient truth for an HIV-positive patient and refusing to believe it does not seem to serve any plausible non-epistemic goal of his. At most, one would expect Ivan to be in denial about there being no safe or effective treatment for his condition (if there wasn’t one), not about there being such a treatment (given that there is one). A much more plausible diagnosis is that Ivan fails to believe in the safety and effectiveness of antiretroviral drugs because he has been exposed to misinformation on the topic of the kind spread by untrustworthy sources such as Joan Shenton’s book Positively False.Footnote ¹⁰⁸

If, as argued in Section 1, the primary concern about science denial is that it is practically harmful, high-stakes cases are likely to raise the most serious practical concerns. This means that, if the arguments in this subsection are correct, the counter-deniers’ practical concerns are largely misplaced. In cases in which someone’s safety or well-being (or that of their loved ones) is at stake, misinformed non-belief seems a much more plausible diagnosis than motivated non-belief.Footnote ¹⁰⁹

Before concluding this subsection, let me mention two important points. The first point is that, while arguments in this subsection suggest that misinformed non-belief is a more plausible diagnosis than motivated non-belief in high-stakes cases, this does not necessarily mean that it is the only or most common diagnosis in such cases. In many high-stakes cases, prudential non-compliance (Section 2.7) might be a better diagnosis. However, the argument is meant to show that either of these two diagnoses is more likely than a motivated non-belief diagnosis in many high-stakes cases. Moreover, even if misinformed non-belief and prudential non-compliance are distinct phenomena, there might be some similarities between them. In fact, in many cases, the difference between the two might be largely a matter of degree. For example, exposure to misinformation might lower someone’s credence in a widely accepted scientific claim enough to turn them into a prudential non-complier but not enough to turn them into a misinformed non-believer. For another, exposure to misinformation might play an important in shaping the preferences of prudential non-compliers (e.g., affecting their perception of the severity of a disease).

The second and, in my view, most important point is that the discussion of misinformed non-belief suggests an alternative diagnosis of the collective epistemic malaise afflicting liberal democracies, which I call the post-trust diagnosis. Unlike the post-truth diagnosis, which claims that the malaise originates from a lack of interest in (or concern for) the truth among a growing number of our fellow citizens, the post-trust diagnosis maintains that it originates from a lack of trust in the institutions forming the backbone of the epistemic infrastructure of liberal democracies on the part of a growing number of communities. The epistemic infrastructure of a society is the system of institutions, norms, and practices that promote the reliable production, transmission, reception, and uptake of information and impede the production, transmission, reception, and uptake of misinformation. The institutions that form the backbone of the epistemic infrastructure of liberal democracies include not only science but also government and the press. According to the post-truth diagnosis, for example, vaccine hesitancy would be due to the fact that (allegedly) a growing number of people have no interest in (or concern for) the truth about the safety of vaccines. However, as argued in this subsection, this explanation seems to be much less plausible than the one offered by the post-trust diagnosis, which maintains that vaccine hesitancy is due to the fact that a growing number of communities do not trust science to conscientiously and objectively investigate the safety of vaccines, government to responsibly regulate their use, or the press to independently monitor those institutions and hold them accountable. This mistrust leaves those communities more vulnerable to misinformation about vaccines (as well as misinformation about those institutions).

3.4 Misinformed Non-Belief in Low-Stakes Cases

The previous subsection argued that, in high-stakes cases, misinformed non-belief and prudential non-compliance are much more plausible diagnoses than motivated non-belief. This subsection argues that misinformed non-belief might be a better diagnosis even in some low-stakes cases that, superficially, might appear to be particularly amenable to an explanation in terms of motivated reasoning One such case is the rejection of the theory of evolution by biblical literalists. Despite appearances, it is far from obvious that their disbelief stems from motivated reasoning aimed at protecting their religious beliefs. Presumably, some creationists are ‘true believers’, who do not engage in motivated reasoning at all. Consider, for example, the case of Karen, a woman who grew up in a Christian community that believes that the Bible is the word of an omniscient and benevolent God and that it is literally true in its entirety. Karen believes that the story of the creation as told by the Bible is the literally true account of the origins of life on Earth and wholly rejects the theory of evolution. Superficially, Karen’s case may seem a prototypical case of motivated reasoning. It seems plausible to assume that, in reasoning about the origins of life, Karen aims to achieve the non-epistemic goals of protecting her religious beliefs or affirming her identity as a member of her religious community at the expense of the epistemic goal of having true beliefs about the topic. However, this does not need to be the case. In fact, for all we know, Karen’s only goals in reasoning about the origins of life could be purely epistemic. Karen knows that most scientists disagree with her and, instead, believe a completely different explanation of the origins of life on Earth. However, the scientists themselves concede that science is fallible and that even their best theories could be false. God, on the other hand, is infallible, or so Karen believes. Given her background beliefs, it seems that, if Karen is exclusively interested in having true beliefs about the origins of life, she should believe the account found in the Bible, not the one widely accepted by scientists. While Karen does reject the theory of evolution because of her religious beliefs, this need not be due to motivated reasoning. In fact, we can assume that her goals are purely epistemic. Karen believes that she has a better chance of achieving her purely epistemic goals by believing the truthful word of an infallible God rather than the defeasible theories of fallible scientists. Once again, if Karen’s conduct is epistemically irresponsible (or blameworthy), it is not because she engages in motivated reasoning, but because she too, like Frank (Section 3.2) and Helen (Section 3.3), trusts sources, such as the Bible and her minister, that are not trustworthy about this specific topic and she mistrusts sources that are more trustworthy on this topic, such as the vast majority of biologists.

Of course, this is merely a fictional example. It is possible that most or all real-world creationists are motivated non-believers rather than misinformed non-believers like Karen. However, Karen’s case still teaches us two important lessons. First, it suggests that, even if all actual creationists are motivated non-believers, they are likely to be ‘impure’ motivated non-believers, as it is undeniable that creationists are typically exposed to large amounts of misinformation about specific scientific topics (as well as about higher-order misinformation about the nature of scientific inquiry and its aims) and implausible to assume that this misinformation has no effect on their beliefs. Second, it shows that even cases that, at first, appear particularly amenable to explanations in terms of motivated reasoning may, on closer scrutiny, turn out to be cases of misinformed non-belief (or, at the very least, cases of ‘impure’ motivated non-belief). In other words, the mere failure to believe an inconvenient truth is not sufficient to establish a motivated non-belief diagnosis. More evidence is needed to avoid misdiagnosing cases of misinformed non-belief (or ‘impure’ motivated non-belief) as ‘pure’ cases of motivated non-belief. In the next subsection, I argue that this evidence is likely to be particularly elusive in low-stakes cases.

3.5 Is Motivated Non-Belief Discernible from Insincere Contestation?

In the previous subsection, I have argued that, in order to reliably discern misinformed non-believers from motivated non-believers in low-stakes cases, more evidence is needed that their beliefs are actually the result of motivated reasoning. In this subsection, I argue that this evidence is likely to be particularly elusive in those cases. In low-stakes cases, it is extremely difficult to distinguish alleged cases of motivated non-belief from cases of insincere contestation. If my arguments are correct, then motivated non-belief diagnoses in low-stakes cases are threatened on both flanks. On one side, as we have seen in the previous subsection, (‘pure’) motivated non-belief might be difficult to discern from misinformed non-belief (or ‘impure’ motivated non-belief). On the other side, it is difficult to discern from insincere contestation.

To see the last point, consider the cases of Ben and Jerry. Ben is a motivated non-believer. He is libertarian who works as an engineer for an oil company, drives a gas-guzzling pickup truck, and supports political candidates who share his views about free markets and small government. Despite his engineering background and the fact that he knows more about climate science than most, Ben is in denial about the reality of anthropogenic climate change because his reasoning about it prioritizes the non-epistemic goals of protecting his political beliefs and of justifying his career and lifestyle over the epistemic goal of having true beliefs about the climate. As a result, Ben sincerely believes that the evidence for anthropogenic climate change is weak and that the climate scientists who claim otherwise do so either as a matter of professional interest (e.g., to receive large grants) or to promote their own political agendas (e.g., they are environmentalists or socialists, who support Big Government and oppose free markets). However, ironically, it is, in fact, him who is trying to protect his economic interests and his political beliefs through motivated reasoning. Jerry, on the other hand, is exactly like Ben except for the fact that, unlike Ben, he is an insincere contester. Jerry does not engage in motivated reasoning on climate change. In fact, he is privately inclined to believe in the reality of anthropogenic climate change. However, he claims not to do so to further his own political and economic interests.

The problem for counter-deniers is that it might be extremely difficult to tell Ben and Jerry apart. In other words, it is extremely difficult to determine whether someone like Ben genuinely does not believe in anthropogenic climate change or (like Jerry) merely claims not to believe in it to serve his own non-epistemic goals (e.g., justify his lifestyle decisions, his opposition to policies with which he disagrees on political or ideological grounds, or even to just protect his well-paid job). For all we know, people who behave like Ben and Jerry might be insincere contesters (like Jerry) rather than motivated non-believers (like Ben).

The problem here is partly that the verbal behaviour of Jerry is likely to be indistinguishable from that of Ben. Talk is cheap. If Jerry were asked whether he really does not believe in anthropogenic climate change, he would insist that he really does not accept the reality of anthropogenic climate change. However, in light of Jerry’s own claims, Jerry’s denials would ring hollow. After all, Jerry (like Ben) would be the first to suggest climate scientists say things that they do not actually believe to serve their own political or professional interests, so why would the same not apply to Jerry?

This worry seems to be particularly troublesome given that one of the main ways to estimate the prevalence of science denial is to conduct surveys (such as the ones run periodically by the Pew Research Center in the United States).Footnote ¹¹⁰ The problem is that, if, as counter-deniers seem to believe, one of the primary goals of motivated non-believers is to affirm their social identity, then it seems plausible that someone like Jerry might affirm his social identity by not responding truthfully to surveys and claiming not to believe in anthropogenic climate change even if he does. This is particularly true if he sees this as a chance to provoke the sort of ‘soy-latte-sipping liberals’ who, in his mind, read these surveys and fret about the rise of climate-change denial. It is likely that most extant surveys miscount insincere contesters as science deniers and, as a result, grossly overestimate the prevalence of science denial (understood as motivated non-belief).Footnote ¹¹¹

Moreover, in low-stakes cases, the non-verbal behaviours of alleged science deniers are also likely to be virtually indiscernible from those of insincere contesters. To see why, compare Jerry’s case to a high-stakes case – that of Logan, a tobacco executive who quit smoking despite publicly downplaying the health risks of tobacco. While Logan’s verbal behaviour might suggest that he is a science non-believer, his non-verbal behaviour is a tell-tale sign of the insincerity of his science contestations. If Logan actually believed his own claims, he would put his lungs where his mouth is. In low-stakes cases (like those of Ben and Jerry), on the other hand, it is very unlikely for there to be any non-verbal tell-tale sign. Most of Ben’s and Jerry’s choices (e.g., what vehicle they drive, what job they have, what political candidates they support) are likely to be the result of a wide variety of factors both within and outside their control and their beliefs about anthropogenic climate change are likely to play a relatively minor role in those decisions and, at any rate, they are very unlikely to be a decisive factor. In other words, it seems implausible to assume that, if Ben were to believe in anthropogenic climate change (despite his science contestations), he would drive a different vehicle than Jerry, have a different job than Jerry, or support different political candidates than Jerry.

These considerations suggest that, despite their different beliefs about climate change, Ben’s and Jerry’s behaviours (both verbal and non-verbal) might be completely indistinguishable from one another. There is no tell-tale sign of the insincerity of Jerry’s science contestation or of the sincerity of Ben’s disbelief. In fact, only Jerry can know that he is an insincere contester. Ben, on the other hand, does not even know he is a motivated non-believer, because, as long as he is successfully in denial about the reality of anthropogenic climate change, he must also be in denial about being in denial.Footnote ¹¹²

3.6 Who Needs to Be a Motivated Non-Believer?

This subsection argues that someone like Ben does not need to engage in motivated reasoning to pursue the non-epistemic goals that are usually attributed to motivated non-believers (and other motivated reasoners). Those goals could be pursued equally well (and, possibly, more cheaply) in other ways. Here, I discuss the two sorts of non-epistemic goals that are most often attributed to motivated non-believers. The first is to affirm their social or political identities. The second is to protect their social, political, or religious beliefs. Note that the arguments in this subsection only target those who, like counter-deniers, adopt a particularist conception of motivated reasoning and apply only to people who are (supposedly) ‘pure’ motivated non-believers (i.e., they are motivated non-believers who have not been exposed to misinformation on the relevant scientific topics).

Let me consider first the goal of affirming one’s social or political identity. We are social animals and, as such, we use all kinds of means (e.g., items of clothing, flags, words, actions) to signal our belonging to certain social groups (e.g., Republicans or Democrats, rural or urban, working-class or middle-class, Evangelical Christian or Hindu, Lakers fan or Celtics fan). According to counter-deniers, some of the science deniers’ beliefs are signals of this kind. For example, Matt might reject anthropogenic climate change in order to affirm his socio-political identity as a rural working-class conservative man. ‘People like him’ don’t believe in anthropogenic climate change! However, Matt does not actually need to actively disbelieve the reality of anthropogenic climate change in order to successfully signal his belonging to a socio-political group. He can do so equally well by faking the signal – that is, by merely claiming to disbelieve the relevant scientific claims. In fact, the latter option would seem cognitively cheaper. I take it that belief is largely involuntary and, since we are assuming that Matt is a ‘pure’ motivated non-believer, he has not been exposed to any misinformation about climate change.Footnote ¹¹³ This means that either Matt has not been exposed to any information about it (and thus has no opinion on it), or he has only been exposed to truthful information about it (and thus he is inclined to believe in anthropogenic climate change). In either case, his disbelieving the reality of anthropogenic climate change would require an effort to believe against the evidence (or lack thereof). This is why Matt would supposedly have to engage in motivated reasoning by, for example, seeking out misinformation that to support disbelief in anthropogenic climate change. However, that requires time and effort. Matt, however, does not need to go through the effort required to believe against the evidence. He could achieve the goal of affirming his social identity equally well (and cognitively more cheaply) by being an insincere contester.

Counter-deniers might reply that self-deception is cognitively cheaper than deception and that people typically deceive themselves in order to better deceive others.Footnote ¹¹⁴ However, for one thing, it is unclear that is generally true, and, for another, even if it is, there is no reason to think that Matt must engage in deception to be an insincere contester. Matt could fake the signal equally well (and more cheaply) by bullshitting about anthropogenic climate change rather than lying about it.Footnote ¹¹⁵ This diagnosis seems to fit particularly well those cases in which people switch between different claims (‘climate change is not happening’, ‘it is happening but it is not caused by human activity’, ‘it is caused by human activity, but it is no cause for alarm’, etc.).

Consider now the goal of protecting one’s social, political, or religious beliefs. Again, it seems that this non-epistemic goal can be served equally well (and cognitively more cheaply) by other means. Suppose that Ben has not been exposed to any misinformation about climate change. Like Matt, he would have to disbelieve the reality of anthropogenic climate change despite the evidence, and like in Matt’s case, this would require him to seek out misinformation about climate change, which requires time and effort. However, no such effort would be needed to square Ben’s political beliefs in free markets and small government with the reality of anthropogenic climate change, as there are plenty of other ways for him to do so. For example, Ben could simply believe that the best response to anthropogenic climate change is to adapt to it rather than trying to mitigate it.

To my mind, the best explanation of why, in the real world, people like Ben tend to take the more arduous route of disputing the reality of anthropogenic climate change despite the scientific consensus about it is that, unlike Ben, they have been exposed to misinformation about it from a variety of sources they trust. The division of epistemic labour applies also to questions of policy. The evidence seems to suggest that most people defer to their favourite politicians, political commentators, or political parties when it comes to deciding what political issues are important and which policies would best address those issues.Footnote ¹¹⁶ This suggests that, if those who have a vested interest in delaying the adoption of policies to mitigate climate change had adopted early on the strategy of advocating for adapting to climate change (instead of the strategy of disputing the scientific status of anthropogenic climate change), it is highly unlikely that people like Ben would have disbelieved the reality of anthropogenic climate change instead of believing in it but disputing that mitigation is the best response to it.Footnote ¹¹⁷

3.7 Is Motivated Non-Belief Practically Harmful?

Earlier (in Section 3.3), I argued that misinformed non-belief is a more plausible diagnosis than motivated non-belief in high-stakes cases of science non-belief and also that this shows that counter-deniers’ practical concerns are largely misplaced, as the high-stakes cases are the ones that are more likely to have practically harmful consequences. Counter-deniers, however, might argue that motivated non-belief can still be practically harmful in low-stakes cases even if the practical harms in these cases are likely to be more indirect and diffuse than in high-stakes cases.Footnote ¹¹⁸ For example, they might argue that, as a result of his motivated non-belief, Ben is more likely to make practically harmful choices in his capacity as a consumer (e.g., buying a fuel-inefficient car, travelling more by plane) or as a citizen (e.g., support political candidates who oppose climate mitigation policies). While the practical harms caused by Ben’s choices might be negligible in the grand scheme of things, they still contribute to harming our planet and everyone on it, including Ben himself.

The main problem with this line of reasoning is that it presupposes that, in cases of motivated reasoning, the relationship between beliefs and actions is the same as in ordinary cases. However, this assumption does not seem to withstand scrutiny. If and when people engage in motivated reasoning, they seem to do so in an attempt to rationalize their pre-existing beliefs, preferences, and decisions.Footnote ¹¹⁹ Consider again the case of Andy. Presumably, the reason why Andy is in denial about his drinking habits is that, if he wasn’t, then, at least, he would have to acknowledge the fact that his drinking habits are negatively affecting him and those around him in various ways, which might make him feel guilty and embarrassed. Moreover, he might have to do something about it, which he might not be willing or able to do, which might make him feel inadequate and powerless. Being in denial allows Andy to keep doing what he does while protecting him from the negative feelings that would be likely to arise if he acknowledged his drinking problem and justify his inaction to himself and to others. In other words, Andy’s motivated reasoning is primarily a source of post hoc rationalizations (i.e., rationalizations of his pre-existing beliefs, preferences, and decisions). But this seems to suggest that, when applied to motivated non-belief, the counter-deniers’ practical concern stems from a misconception of the relationship between beliefs and decisions in motivated reasoning.

Our practical decisions as consumers and as citizens are typically the result of a wide variety of factors both within and outside our control and, especially in low-stakes cases, it seems very unlikely that our beliefs about scientific claims play a significant role in those decisions and even more unlikely that they play a decisive role – that is, that if we were able to change the minds of motivated non-believers, they would be making different practical decisions. Consider, for example, the case of air travel. Anecdotally, the vast majority of the academics I know (myself included) believe in anthropogenic climate change, and yet, even if it is well-known that air travel is one of the main contributions individuals can make to global carbon emissions, most academics I know (including, admittedly, myself) do not drastically limit their work-related air travel. Systematic empirical studies seem to support my observations (for both academics and non-academics).Footnote ¹²⁰ In fact, a recent study found that climate scientists fly more for work than other academics.Footnote ¹²¹

This in itself is a major problem for the argument we are considering, as it suggests that, even if Ben is indeed a motivated non-believer, his behaviour would not change significantly if only he started believing in anthropogenic climate change. In particular, it is unlikely that, even if we could magically change his mind about anthropogenic climate change, he would buy a different car, get a different job, or support different political candidates. Ultimately, whether or not, deep down, Ben believes in anthropogenic climate change seems to be largely irrelevant to his behaviour as a citizen or as a consumer, as, in low-stakes cases, there are plenty of other potential rationalizations Ben’s could employ to justify (or excuse) his pre-existing practically harmful decisions to himself or to other.

It is important to note that the arguments in this subsection are not meant to establish that motivated non-belief is completely practically harmless. For one thing, it is plausible to suppose that motivated reasoning strengthens a person’s confidence in their own beliefs, preferences, and decisions. Their upshot, in my view, is that, even if motivated reasoning has these effects, there is still little or no reason to believe that it ever plays a decisive role in practical decisions in low-stakes cases, as, arguably, these are affected by a wide variety of much weightier factors, such as social and economic factors. If so, then, in low-stakes cases, motivated non-belief is, at most, a fig leaf to cover up one’s true motivations. If this is correct, then the counter-deniers’ practical concerns would be misplaced even in genuine cases of motivated non-belief.

3.8 What to Do about Motivated Non-Belief?

Even if my arguments so far are wrong and motivated non-belief makes a decisive contribution to practically harmful decisions, counter-deniers still face a dilemma. On the one hand, insofar as motivated non-believers’ decisions are practically harmful to others, it is unclear that we should solely try to persuade them to make different decisions. After all, we do not merely try to persuade smokers not to expose others to the practically harmful consequences of their smoking – we forbid them to smoke in public places. One of the central tenets of liberalism is that people should be free to make their own decisions but only as long as those decisions do not harm others. On the other hand, if motivated non-believers’ decisions are exclusively harmful to them, then it is unclear that there is anything we should do about it (other than, possibly, trying to convince them not to make those decisions by means of rational persuasion).Footnote ¹²² Another central tenet of liberalism is its anti-paternalism – competent adults should be free to make their own decisions even if those decisions might have practically harmful consequences for them (as long as they don’t harm anyone else). For example, even if smoking does have practically harmful consequences for smokers, we do not forbid smokers from smoking in their own homes. At most, we try to persuade them to quit smoking.

Arguably, the best response to this dilemma is to try to blunt its horns. The first part of this response aims to blunt its first horn. While it is possible to limit the freedom of people when their actions might harm others, there must be some proportionality between the harm those decisions would cause to others and the harm caused to them by curtailing their freedom to make those decisions. For example, even if the seasonal flu kills people every year, we collectively judge that the harms of limiting the freedom of flu patients by quarantining them would outweigh the harms caused to the rest of society. This, however, does not mean that we should not try to encourage flu patients to take steps to avoid spreading the flu. In other words, the short reply to the first horn of the dilemma is that, in many cases, there is room for non-coercive interventions. The second part is that, while liberalism opposes paternalism in theory, liberal democracies often adopt paternalistic policies in practice. The standard example of such policies are laws enforcing the use of seatbelts. Again, the issue seems to be one of proportionality. We collectively judge that the practical harms caused by letting people freely decide whether to wear a seatbelt far outweigh the harms caused by the minimal invasion of freedom that is caused by imposing the use of seatbelts.

While this response blunts the horns of the dilemma, it does so at the cost of raising the bar for counter-deniers. Counter-deniers must not only show that motivated non-belief is practically harmful (despite the arguments to the contrary discussed earlier) but also that the practical harms it supposedly causes fall within the grey area between what can be legitimately regulated by a liberal democracy and what cannot. However, once we look more closely at cases of (alleged) motivated non-belief, it is unclear that they typically fall within grey area. Suppose, for example, that, contrary to my arguments, Ben’s motivated non-belief is a decisive factor in his practically harmful decisions as a consumer (e.g., which vehicle to purchase) or as a citizens (e.g., which political candidate to support). Neither kind of decisions seems to fall into the grey area between what can be legitimately regulated by a liberal democratic government and what cannot. On the one hand, Ben’s decision to buy a highly fuel-inefficient vehicle is practically harmful not only to him but also to everyone else and, in fact, to the whole planet. It is, therefore, a clear instance of the kind of decision that can be legitimately restricted in a liberal democracy (by, e.g., increasing fuel-efficiency standards). On the other hand, his decision to support a political candidate is an instance of the kind of decision that a liberal democracy cannot legitimately restrict (except in extreme cases), as a political system that does not allow people to support their preferred political candidate is not truly a democracy, let alone a liberal one.Footnote ¹²³

3.9 Conclusion

In this section, I have argued that, despite the promise of the motivational account, counter-deniers face a dilemma. On the one hand, while motivated non-belief might be always epistemically irresponsible (or blameworthy), it is dubious that it is ever practically harmful. On the other hand, while misinformed non-belief may sometimes be practically harmful, it is doubtful that it is always (or ever) epistemically irresponsible (or blameworthy).

More specifically, I have argued:

• that misinformed non-belief is a much more plausible diagnosis than motivated non-belief in high-stakes cases, such as vaccine hesitancy or HIV/AIDS denial,

• that misinformed science non-belief might be the correct diagnoses even in low-stakes cases that superficially appear particularly amenable to an explanation in terms of motivated reasoning, such as creationism,

• that, in low-stakes cases, motivated non-belief might often be indiscernible from insincere contestation, which, among other things, makes it difficult to determine the actual prevalence of motivated non-belief,

• that there are cognitively cheaper ways to achieve the non-epistemic goals usually attributed to motivated non-believers,

• that it is unclear that, in low-stakes cases, motivated non-belief is as practically harmful as counter-deniers assume, as it might be just a way for the alleged motivated non-believers to rationalize their own pre-existing beliefs, preferences, and decisions rather than a decisive factor in those decisions, and

• that, even if motivated non-belief is, in fact, practically harmful, its practical harms are either of the kind that may be prevented through the legitimate use of coercion or of the kind that cannot be prevented without undermining the democratic nature of the political system.

While much more could be said about each of these theses, I think that taken together they form a strong presumptive case against the view that science denial (understood as motivated non-belief) warrants both the epistemic concerns and the practical concerns of counter-deniers. I take it that there are a few lessons to be learned from this discussion.

The first lesson is that, if science denial is identified with motivated non-belief, then it becomes exceedingly difficult to distinguish it from a number of superficially similar phenomena, such as misinformed non-belief, insincere contestation, or prudential non-compliance. The narrative that science denial is on the rise is supported primarily by three types of empirical evidence: anecdotal evidence, direct systematic evidence (e.g., opinion surveys), and indirect systematic evidence (e.g., evidence about vaccination rates). However, once those specific phenomena are distinguished from one another, it becomes apparent these bodies of evidence do not give us a reliable way to differentiate among them. People seem to have all kinds of reasons to engage in science contestation in certain conversational contexts (such as opinion surveys) and, even more so, to act as they do in a variety of practical contexts (such as when making decisions about vaccinations). I think the lesson here is that we should try to better understand the beliefs and concerns of each alleged science denier by actually engaging with them (rather than reproaching them from afar as counter-deniers are wont to do).

The second lesson is that, if science denial is identified with motivated non-belief, then it is doubtful that it is ever practically harmful, as it is unlikely to be a decisive factor in the practically harmful decisions of the (alleged) motivated non-believer. It is much more likely that motivate reasoning helps them produce post hoc rationalizations of their practical decisions.

The third lesson is that, unlike motivated non-belief, misinformed non-belief and prudential non-compliance are likely to be practically harmful in high-stakes cases (such as those of childhood vaccinations, COVID-19, or HIV/AIDS). However, neither is plausibly construed as a form of science denial. For one thing, it is doubtful that either is always epistemically irresponsible or blameworthy, and, even when they are, the epistemic irresponsibility involved is different both in degree and in kind from the kind of epistemic irresponsibility allegedly involved in motivated non-belief. For another, insofar as motivated non-believers are epistemically irresponsible, it seems to be because they place more epistemic trust in epistemically unreliable sources than in epistemically trustworthy ones. What I think we should learn from this is that we need much finer-grained categories than ‘science denial’ to be able to properly identify the sources of potentially harmful behaviour in high-stakes cases and develop effective approaches to addressing them appropriately and respectfully. For example, we need a better appreciation of the fact that vaccine hesitancy is a complex and multifaceted phenomenon and that to brand all vaccine hesitators as science deniers is likely not only inaccurate but also counterproductive.Footnote ¹²⁴ A much more promising approach would be to start from trying to understand vaccine hesitators’ concerns in an empathetic and non-judgmental manner and then trying to address them.Footnote ¹²⁵

The fourth and final lesson is that science contestation (whether sincere or insincere) is likely to be epistemically harmful, as it pollutes the epistemic environment with the kind of misinformation that seem to feed misinformed non-belief and certain forms of prudential non-compliance. In fact, if we do not want to abandon the pejorative term ‘science denial’, science contestation (especially in its insincere form) is arguably the phenomenon that is most deserving of it. However, if the post-trust diagnosis is correct, then the most promising approach to preventing science contestation from feeding misinformed non-belief and other practically harmful behaviours is not to try to prevent individuals from being exposed to misinformation (which, in a liberal democracy, is neither feasible nor desirable) but to immunize communities from misinformation by restoring their trust in the institutions that underpin the epistemic infrastructure of liberal democracies. As I mention in the Conclusion, however, this is easier said than done, as their mistrust is not always entirely unjustified or misplaced.