On the Uses and Abuses of Celebrity Epistemic Power

ABSTRACT

The testimonies of celebrities affect the lives of their many followers who pay attention to what they say. This gives celebrities a high degree of epistemic power, which has come under scrutiny during the COVID-19 pandemic. This paper investigates the duties that arise from this power. We argue that celebrities have a negative duty of testimonial justice not to undermine trust in authoritative sources by spreading misinformation or directing attention to untrustworthy sources. Moreover, celebrities have a general imperfect duty to try to correct for an unjust distribution of attention by redirecting it to those who deserve it. During a pandemic this may become a perfect one, due to the harm that could be prevented if people follow the advice of experts. Relatedly, we argue that celebrities have an imperfect duty to promote behavior that will reduce the spread of a pandemic. We outline three ways they might do so: they might take on the position of a role model, they may act as a salience magnet or they can direct people’s attention towards others who have taken on these roles.

KEYWORDS:

• Celebrity
• epistemic power
• epistemic justice
• trust

1. Introduction

Miranda Fricker defines epistemic injustice in terms of the wrongs people suffer in their capacity as knowers. In her eponymous 2007 book on this topic, Fricker distinguishes testimonial injustice from hermeneutical injustice. Testimonial injustice – which still receives the lion’s share of philosophical attention – occurs when someone faces systematic credibility deficits, especially when those deficits arise because of identity-based prejudice. Victims of testimonial injustice are trusted less than they deserve to be trusted because of their gender, race, class, or other identity-attribute. By contrast, hermeneutical injustice relates to deficits and distortions in the conceptual and linguistic resources available to people, which can make it difficult for them to explain their experiences and attitudes to others or even to understand them themselves. Those who have gone on to develop Fricker’s insights taxonomized various forms that testimonial injustice can take. For instance, Medina (2013) emphasizes injustices that arise from credibility excesses (as opposed to deficits) that accrue to people who belong to dominant groups. In the same vein, Hookway (2010) explores how epistemic injustice occurs when a person fails to take another person’s questions (as opposed to their assertions) as seriously as they deserve. And Nikolaidis (2021) contemplates what he calls formative epistemic injustice, which occurs when someone’s epistemic capacities are malformed due to, for example, racist educational practices (e.g. not allowing slaves to learn to read). Such contributions to the taxonomy of epistemic injustice have practical implications for how we understand oppression, silencing, and privilege.

We explore another kind of testimonial injustice, one that currently receives remarkably little attention. We are interested in how certain individuals are embedded within structured testimonial networks that enable them to command much larger audiences than their assertions often deserve.¹ Thus we move beyond existing approaches to epistemic injustice, which typically focus on dyadic exchanges between a single speaker and a single hearer. When Medina, Hookway, and other epistemic injustice scholars follow Fricker in speaking of credibility deficits and excesses, they have in mind the amount of credence that a hearer places in the assertions of a speaker. While this is also our starting point, we only discuss it insofar as it informs non-dyadic situations in which asserters have inordinate epistemic power due to their vast audience of fans or followers.² We can often observe positive correlations between the credibility of a speaker and the size of their audience. Some people are highly trustworthy, highly trusted, deserve large audiences, and have large audiences. Others are not at all trustworthy, not much trusted, deserve at most a small audience, and have a small audience. Nevertheless, these connections can – and with the help of digital technologies increasingly do – come apart. On the one hand, there are Cassandras who are ignored despite legitimately warning against grave future challenges (e.g. climate change). These individuals deserve to be trusted but only have modest or skeptical audiences. On the other hand, there are blowhards who use their gigantic platforms to, for instance, promote conspiracy theories. Such individuals reach millions, even though they deserve to be distrusted and ignored.

To explore the consequences of these phenomena, we discuss the duties of celebrities during the COVID-19 crisis. We begin by introducing the epistemic power possessed by celebrities and in particular the impact this power can have during a pandemic (Section 2). We then explore the moral responsibilities that may arise from this power (Sections 3–6). We will argue that celebrities have a negative duty of testimonial justice not to undermine trust in authoritative sources by spreading misinformation or directing attention to untrustworthy sources. Moreover, celebrities have a general imperfect duty to try to correct for an unjust distribution of attention by redirecting it towards those who deserve it. During a pandemic this duty may become a perfect one, due to the harm that could be prevented if people follow the advice of experts. Relatedly, we argue that celebrities have an imperfect duty to promote behavior that will reduce the spread of a pandemic. We outline three ways they might do so: they might take on the position of a role model, they may act as salience magnets or they can direct people’s attention towards others who have taken on these roles.

2. Celebrities in the Age of COVID-19

The COVID-19 crisis presents a wide array of challenges, both those that must be addressed individually (handwashing) and those that require a collective response (treatment of the sick, limits on panic buying, international coordination, etc.).³ However, as we mentioned above, not all individuals have equal epistemic standing or power. To grasp the epistemic and ethical challenges for celebrities and other prominent individuals in responding to pandemics, we must understand these differences.⁴

A celebrity is someone who is known beyond those they have direct contact with. In Antoine Lilti’s words, the ‘celebrated individual is not known simply to his family, his colleagues, his neighbours, his peers or his customers, but to a vast group of people with whom he has no direct contact, who have never met him and will never meet him, but who frequently encounter his public image’ (2017, 6). Celebrities may be known for a particular skill or achievement, but this is not a necessary feature of celebrity, as they may be famous simply for being famous, as captured by Daniel Boorstin’s claim that celebrities are people ‘well-known for their well-knownness’ (1962, 57). Because celebrities are so widely known, they are the focus of high levels of public attention, and this attention typically extends beyond the particular reason for their fame (van Krieken 2012, 10). The public do not only want to hear about a celebrity’s particular talents but also about their private lives as well as their values, beliefs, and opinions on topics unrelated to their talents.

Celebrities are often also the subject of widespread admiration (Meng-Lewis et al. 2021). People admire Roger Federer’s tennis skills, Beyoncé’s musical talents, and George Clooney’s acting. This admiration is not limited to these talents and often extends into their behavior in their personal lives and their values and opinions. In addition, celebrities are often also the target of para-social relationships (Horton and Wohl 1956). These are one-sided relationships in which fans relate to their idol in the way they would view a friend or family member. They may think and talk about their idol ‘as a “friend”, “older sister”, “father figure”, “guide” or “mentor”’ (Caughey 1984, 53). They may also incorporate the celebrity’s values and plans into their own self-conception in the same way they would with a friend or lover.⁵

Long before the COVID-19 crisis started, celebrities had increasingly become part of an epistemically privileged group. Mass-communication technologies are able to amplify their practical and epistemic power to unprecedented levels, and increasing numbers of fans and followers view them as ethical or epistemic authorities. Indeed, recent papers estimate that the Gini coefficient of the online attention economy, as measured by degree and by PageRank, is approximately 0.90 (Lopes et al. 2011; van Mierlo, Hyatt, and Ching 2016), including the attention economy specifically related to COVID-19 (Quintana et al. forthcoming). This means that a very small number of social-media accounts enjoy the vast majority of online attention, lending their controllers massive agenda-setting powers.

These powers can be used for good or ill. Since the pandemic struck, some celebrities and policy makers have spread medical misinformation. More positively, misuse of ethical and epistemic authority has been counterbalanced by those who have used social media to influence others in ways that are aimed at thwarting the spread of the virus. Encouragement to self-isolate, as well as tips on quarantining, illustrate celebrities using their power pro-socially. Examples abound, including musicians from Britney Spears to Madonna and movie stars from George Clooney to Gwyneth Paltrow. Take Arnold Schwarzenegger’s Instagram posts about the merits of social isolation, for instance. In a short video clip taken in 2020, Schwarzenegger recorded himself smoking one of his trademark cigars and explaining the collective merits of individual self-isolation. By presenting an attractive vision of what quarantine can look like, celebrities hope that their fans will emulate their socially responsible behavior. In addition to providing a potent practical example, celebrities such as Schwarzenegger have used their epistemic power to publicly debunk conspiracy theories and to redirect their followers to reputable medical advice. Given the power of today’s celebrities, such epistemic responsibility stands to have far-reaching consequences for the health and welfare of millions of people. At the time of writing this paper, Schwarzenegger’s social isolation clip had been viewed a mind-boggling 13.6 million times.

Celebrities’ ability to promote public goods and public health has long been recognized. Elvis Presley endorsed the polio vaccination and was vaccinated on live television in 1956. This event has been credited as an important factor in the immunization of adolescents and teenagers, a demographic that had previously shown lackluster uptake of the vaccine (Colgrove 2006, 126–127; Trebach, Soppet, and Sozer 2011, 2).

Despite examples of celebrities using their epistemic authority for good, there are detractors. Take Liverpool’s celebrity football manager, Jürgen Klopp, as an example. Responding to a reporter’s question about his view on the spread of COVID-19, Klopp scolded the journalist, responding that, concerning important matters such as medicine and public health:

What I don’t like in life is that [for] a very serious thing, a football manager’s opinion is important. It’s not important what famous people say. People with knowledge will talk about it and tell people to do this, do that, and everything will be fine, or not. Not football managers. (Zirin 2020)

Klopp objected to the inordinate epistemic power that celebrities now wield, especially at times when their uninformed pronouncements could have a profound effect on others. Klopp’s remarks are salient in a world where celebrities are actively encouraged to share their views, even at times when these opinions are manifestly outside their area of expertise.⁶ Communicating online allows celebrities to share their views and opinions with their fanbase at the click of an online device. Such swift, high-volume, real-time communication is further exacerbated by the fact that online platforms have now, whether intentionally or not, effectively bypassed traditional gatekeepers. As long as one follows the rules on nudity, copyright infringement, and (sometimes) hate speech, there are no gatekeepers to becoming a YouTube star. More traditional credentials of expertise have been replaced with metrics that count the number of views or subscribers that a person manages to attract.

3. The Epistemic Power of Celebrities and Moral Responsibility

Given the often-undeserved epistemic power of celebrities, we should ask whether their influence gives them special duties, particularly during a pandemic or similar kind of public health crisis. This discussion falls under the rubric of testimonial injustice because celebrities operate in networks of attention and trust that, as mentioned above, embody stark distributive injustices, and those distributive epistemic injustices can in turn lead to practical harms such as illness, suffering, and death.⁷

Before addressing our question directly, we investigate three ways in which it can be argued that celebrities are subject to special duties. We will argue that these duties are limited in an everyday context, but that they become more demanding under crisis conditions. During a pandemic such as COVID-19, we contend that there is good reason to think that the epistemic power of celebrities places special negative and positive duties on them.

3.1. Role Model Duties

The most prominent argument for the claim that celebrities are subject to special moral duties holds that they are subject to role model obligations. A role model obligation is an obligation to serve as a good role model, or a moral exemplar, to other people. As Wellman (2003, 335) points out, the prominence of celebrities gives them an unusually high degree of influence over other people.⁸ While everyone might be subject to some kind of obligation to set a good example for others, the wide-ranging influence of celebrities gives us good reason to view them as subject to special duties to serve as good role models. In other words, celebrities have a duty to encourage people to become virtuous rather than vicious (Wellman 2003, 335).

There are two ways to support the claim that celebrities have an outsized influence. First, the simple fact that celebrities receive more attention than ordinary people users means celebrities have a far greater degree of influence over other people’s behavior. Second, celebrities are often, though not always, widely admired. This latter fact is relevant because there is good reason to think that admiring someone often leads to a desire to emulate them. According to Zagzebski (2015, 209), a desire to emulate should be seen as a ‘fundamental feature’ of admiration. We do not have to accept this conceptual claim, though, to hold the more modest generalization that admiration often leads to a desire to emulate because this claim is routinely supported by many social psychological studies (e.g. Algoe and Haidt 2009; Aquino, McFerran, and Laven 2011; Immordino-Yang et al. 2009; Vianello, Galliani, and Haidt 2010).⁹

3.2. Epistemic Power

The second kind of argument that might be made to support the claim that celebrities are subject to special obligations is that celebrities possess a high level of epistemic power, the power to influence what others believe and know (Geuskens 2018; Archer et al. 2020).¹⁰ As the old saying from Spiderman comics goes, with great power comes great responsibility. Celebrities ought not to misuse their epistemic power, which means that, among other things, they should conduct some research on a topic before speaking publicly about it. Again, this power comes from being admired and from being the focus of a great deal of unregulated attention and trust.

Previous work has shown that admiration has an important role to play in influencing how credible celebrities are taken to be (Archer and Matheson 2019). The idea that celebrities are viewed as credible sources seems to underlie the use of celebrity endorsement in advertising and politics, and over-imitation of successful exemplars in causally opaque environments is not only common but also arguably rational (Mercier 2020; Levy and Alfano 2019). It is worth noting, though, that the effectiveness of celebrity endorsements depends on a number of different factors, including the gender and profession of the celebrity, as well as their ‘fit’ with the product they are endorsing (Knoll and Matthes 2017).

Willingness to attribute credibility to celebrities is not limited to decisions about which product to consume. For example, an observational study by Desai and Jena (2016) found a 64% rise in genetic testing for breast cancer amongst American women in the fortnight following the publication of an article by Angelina Jolie about her own experience of taking this test. Similarly, celebrity endorsements have been found to play a role in influencing political opinion and voting intentions (Jackson and Darrow 2005; Veer et al. 2010; Harvey 2017). Again, the effectiveness of these endorsements depends on a number of factors. In a study of UK voters, celebrity endorsements were found to be particularly effective amongst those who do not spend a lot of time thinking about politics (Veer et al. 2010). According to another study, such endorsements are particularly effective for those who are fans of the celebrity concerned (Jackson and Darrow 2005). The important point for our purposes is that celebrities are viewed as more credible than others, although this connection between celebrity and credibility is moderated by a number of variables.

The other important source of epistemic power of celebrities is attention. As Archer et al. (2020) argue, being the focus of attention is a significant source of epistemic power. First, attention provides a means by which testimony can be heard, so it influences what others believe and know. Second, being the center of attention means that celebrities can direct attention to the testimony of others, making them conduits of testimony rather than sources. Finally, attention provides celebrities with the power to set the agenda and determine which issues people think are important. This is especially true in the age of social media, when celebrities can address the public directly, without having to go through intermediaries such as journalists and editors associated with traditional media.

3.3. Salience Magnets as Solutions to Coordination Problems

A third way to argue for special duties of celebrities concerns the social structures that they operate within. In order to solve coordination problems and foster cooperation, it is often valuable to have salience magnets that reliably garner joint attention. A salience magnet, as we use the term in this paper, is an object, person, or ritual that – despite potentially having nothing to do with reasons or rationality – has a propensity to capture people’s attention. Schelling (1960; see also Lewis 1969, Skyrms 1996; Cubitt 2003) famously employed the notion of ‘salience’ to help explain how people sometimes manage to settle on conventions and other game-theoretic equilibrium strategies when they might instead have settled on a different set of strategies or failed to cooperate entirely. According to Lewis, an equilibrium is salient when it ‘stands out from the rest by its uniqueness in some conspicuous respect’ (1969, 35). For instance, a celebrity could publicly and conspicuously make a show of self-quarantining, which in turn would make the equilibrium in which enough of the population self-quarantines salient to a large enough subset of the population.

In public-health crises, coordinated and cooperative action across large populations is paramount. Achieving this can be done in various ways. In small groups that regularly encounter one another face-to-face, there are established mechanisms such as eye contact, simultaneous performance of embodied rituals (e.g. dancing, singing, clapping), and the use of ritual objects or persons who serve as salience magnets and ensure joint attention (e.g. the priest and the eucharist, the speaker at the podium, the fitness instructor in front of the class).¹¹

As the size of groups increases and contact between members grows more remote, distant, and anonymous, many of these mechanisms require modifying or updating if they are to remain effective. Take an everyday example: clock towers that chime on the hour in medieval towns ensure that everyone is in a position to know the time, as well as ensuring that everyone is in a position to know that everyone else is in a position to know the time. This fosters coordination for meetings and other activities. Likewise, a call to prayer that comes from a mosque ensures that all Muslims within earshot know that it is time to engage in a religious ritual, and that all other nearby Muslims are in a position to know this, making collective religious observance possible. In a world that coordinates itself beyond national boundaries and time zones, the value of salience magnets is more important than ever.

What is essential in all of these examples is that there is a technique to ensure that the descriptive and normative conditions for conditional norm-following are in place (Bicchieri 2005, 2016). As Bicchieri and her colleagues show using multiple methods and across a wide range of cases, a large proportion of most populations (usually not quite a majority) are conditional norm followers, in the sense that they will comply with a norm if and only if they both expect enough other people to comply with it and they think that enough other people expect them to comply with it. One way to assure oneself that enough other people are complying is to observe it directly, but in large, disconnected populations this can be difficult. Another way to ensure that enough other people are complying is to receive indirect evidence that they are aware of the norm, which is where salience magnets are relevant. Likewise, one way to assure oneself that the required number of others expect one to comply with a norm is to ask them, but in large, disconnected populations, this can be difficult. Another way to ensure that enough others expect one to comply is to receive indirect evidence about people’s normative expectations.

Once again, this is where salience magnets are useful. For instance, Bicchieri (2016, 164) describes how ‘trendsetters’ (a type of salience magnet) who aim to establish a new norm or undermine an existing norm tend to be effective only when ‘news of their deviance’ spreads. Bicchieri (2016, 172) further argues that trendsetters need to be relatively risk-insensitive, at least when it comes to social sanction, lest they refrain from acting out of conformism or fear. Once again, celebrities often seem to enjoy partial immunity to social sanction; their reputations may suffer temporarily, but in many cases they return to fame within months or years.

The usefulness of salience magnets in bringing about compliance with social norms furnishes another way to argue for the special moral responsibilities of celebrities. The fact that these celebrities are the focus of joint attention allows them to play a role as salience magnets that can provide indirect evidence about how people in general react to these norms. Dyer (1986, 18) makes a similar point when he claims that ‘Stars represent typical ways of behaving, feeling and thinking in contemporary society’. In other words, by publicly complying with social norms, celebrities can help to ensure that other people also comply by sending the message that this is normal and typical behavior.

4. Limits to Celebrity Responsibilities

In the previous section we showed that celebrities might be thought to have special moral duties, but it’s important to acknowledge that various arguments have been raised against this idea.¹² To do this, we respond to the two arguments that we take to be the most plausible objections to the claim that celebrities are subject to special moral obligations.

According to both Spurgin (2012) and Feezell (2005), considerations of privacy limit the extent to which we can assign role model obligations to celebrities.¹³ Privacy is valuable according to Spurgin because it is needed for people to be able to live their lives in the way that they choose.¹⁴ To insist that celebrities be moral exemplars and to publicly share their exemplary deeds with the public is to deny celebrities of this privacy. If we accept this point, we then face the question of under what conditions it is appropriate to attribute role model obligations to people. According to Spurgin (2012, 121), these obligations are acquired by those who take on roles that involve role model duties. For example, the role of a school teacher might plausibly be thought to involve duties to be a role model for the children in their class. Similarly, parents might also be thought to have duties to be role models to their children. In these cases the role model obligations are limited to particular groups of people, the parent’s children and the children in the teacher’s class.

Nevertheless, Spurgin (2012, 124) points out that one can acquire more general role model obligations if one puts oneself forward as a role model. Most obviously, those who claim that they want to be a role model, or who promote themselves as possessing virtues worthy of emulation, or claim to be exemplary characters, plausibly acquire role model duties as a result. According to Spurgin, celebrities who use their public profile to tell others how to behave can reasonably be held to role model duties as a result.

In relation to the possible duties stemming from epistemic power, we might worry that requiring celebrities to speak only about topics on which they have conducted adequate research would place unjustified limits on their freedom of speech. Why, we might wonder, should celebrities be prevented from sharing their uninformed opinions when non-famous people can blamelessly do so? This might seem to be an unreasonable restriction on freedom of speech, especially in the rapidly expanding online world in which posting or tweeting personal views is regarded as an important part of self-expression. John Stuart Mill claimed that people ought to have ‘absolute freedom of opinion and sentiment on all subjects, practical or speculative, scientific, moral or theological’ (1978, 11) and that the freedom to express these opinions is ‘practically inseparable’ (1978, 12) from freedom of opinion. Those impressed by this view are likely to view the idea that celebrities have special duties relating to their epistemic power to be an unreasonable restriction on their liberty.

Nevertheless, there are compelling reasons for not accepting this argument uncritically. Conceding that celebrities have a duty to only speak about topics on which they have conducted adequate research does not commit us to silencing celebrities who violate this duty by, for instance, removing them from social media platforms. Instead, we might think that this is an area where celebrities have the right to do wrong (Waldron 1981). This would mean that, while celebrities would be wrong to violate this duty, it would also be wrong for the state to enforce compliance. Even viewing the duty as one that celebrities ought to have a legal right to violate may still seem unreasonably burdensome for celebrities, however. Is it always wrong for celebrities to share their opinion on topics they know little about? How can we resolve this conundrum?

5. Negative Duties of Celebrities

During a public health crisis such as the recent pandemic, the primary duty that celebrities face is the duty not to spread misinformation about the crisis. As noted above, one might be reluctant to claim that celebrities face a general duty to act as role models or to only speak publicly on topics they know something about. This could threaten or curtail much that is valuable about self-expression. In this section, however, we provide what we view as a better argument for the claim that celebrities can have special duties not to spread misinformation that arise from the power of their speech to cause or exacerbate testimonial injustice. These duties are particularly strong during a pandemic when testimonial injustices against those with domain expertise (e.g. epidemiologists) can have a severely negative impact on our ability to effectively respond.

The first reason to accept this more specific duty is that the epistemic power possessed by celebrities means that their testimony will be widely publicized. The COVID-19 crisis has highlighted the importance of the epistemic power possessed by celebrities. Celebrities such as Woody Harrelson and John Cusack have been criticized for using Instagram and Twitter to spread the unsupported claim that 5G radio antennas are responsible for the spread of the virus (Perrigo 2020). Brennan et al. (2020) conducted an analysis of 225 items of misinformation related to COVID-19 published between January and March of 2020 that fact-checkers had rated as either false or misleading. They found that while politicians, celebrities, and other public figures produced or spread 20% of the claims in their sample, they nevertheless received 69% of the total social media engagement. They conclude that ‘prominent public figures continue to play an outsized role in spreading misinformation about COVID-19’. This is especially troubling given that a recent study of UK adults found that those who accepted one or more of the three most prominent conspiracy theories about COVID-19 were less likely to comply with public health guidelines designed to minimize the spread of the virus (Allington and Dhavan 2020). Because of the rampant spread of this kind of misinformation by celebrities, the testimony of public health experts will often be given less attention and trust than it deserves.

During a pandemic, the reasons against prominent public figures spreading misinformation are especially strong due to the potential for practical (as opposed to merely epistemic) harm. In some cases, the link between misinformation and harm is obvious. In one tragic case, a man from Arizona died after drinking chloroquine phosphate in an attempt to protect himself from COVID-19. According to his wife, he decided to drink the liquid after they watched Donald Trump promote it as a treatment on NBC News. The video clip of Trump’s claim was shared by many of his 86 million Twitter followers, which had the effect of amplifying viewings of the news segment on a vast scale (Edwards and Hillyard 2020). While this is an extreme case, it highlights how quickly misinformation can spread online. This indicates that celebrities (including celebrity politicians like Trump) who spread misinformation may be causing significant damage to public health.

Similar problems have arisen in relation to the spread of misinformation about vaccinations against COVID-19. For example, rapper and singer Nikki Minaj, who has more than 22 million followers, was strongly criticized for a tweet that claimed that her cousin was left by his fiancé after experiencing testicle swelling as a result of the COVID-19 vaccine. Political commentator Joy Reid claimed that Minaj had ‘put people in the position of dying from a disease they don’t have to die from’ (Olutola 2021). Similarly, basketball star Kyrie Irving’s public explanations of his decision to remain unvaccinated have been criticized by many. This criticism has even been extended to other players, like Lebron James, who are vaccinated but refuse to criticize the anti-vaccination statements of their fellow players. Former NBA star Kareem Abdul-Jabbar, for example, claimed that James’ public statements expressing a desire to honor the individual wishes of players who wish to remain unvaccinated was ‘irresponsible’ (Abdul-Jabbar 2021). Celebrities can also bring about testimonial injustices and harms as knock-on effects by directing trust towards other people who spread misinformation. This can be done either by retweeting and thus amplifying them or by direct endorsement. The latter case is illustrated by the recent behavior of Phil McGraw, a television personality who hosts talk show ‘Dr. Phil’. Despite lacking any expertise in epidemiology, he has continually called for COVID-19 lockdown restrictions to be lifted. At the first-order level, we can say that McGraw is using his influence in a way that is dangerous for public health. At the second-order level, McGraw has been promoted, retweeted, and directly endorsed by other celebrities, including Oprah Winfrey, who has a staggering 43 million Twitter followers. As fellow celebrity Seth MacFarlane puts it, ‘Oprah has done some wonderfully altruistic things with her career, but the use of her platform to amplify the voices of dubious characters rather than legitimate scientists has been a disservice’ (Wilstein 2020). Celebrities who direct attention to other celebrities such as McGraw can undermine the credibility of public health experts and cause significant harm by amplifying McGraw, legitimizing his testimony, and directing trust towards him.

Given the potential testimonial injustices and harms that can be caused by the spreading of misinformation online or misdirecting trust during pandemics by those with epistemic power, celebrities and other public figures have a duty not to do this. This conclusion is supported first by a basic duty not to act unjustly or to act in ways that perpetuate injustice. Just as those who fail to believe someone they ought to believe violate a duty to make a proper testimonial judgement (Fricker 2007, 26), so too do those who undermine the credibility of other people in ways that lead other people to violate their duty to make a proper testimonial judgement.

This conclusion is also supported by the general prima facie duty to avoid harming others, otherwise known as the principle of non-maleficence. This is a widely accepted ethical principle featuring in Ross’s (1930) list of seven prima facie duties and Beauchamp and Childress’s (1977) list of four key principles of biomedical ethics. Indeed, this duty is so widely accepted that we have good reason to be skeptical of any moral outlook that cannot accommodate it.

It does not follow straightforwardly from the claim that celebrities may act wrongly when they spread misinformation that it would be right to legally restrict celebrities’ ability to speak freely on these topics. The fact that celebrities may be justifiably criticized for such speech does not by itself justify imposing restrictions on their speech. However, it is worth noting that the prohibition on causing harm is often thought to justify restrictions on free speech. Even those who are willing to accept very few restrictions on freedom of speech tend to accept that speech that predictably causes harm should be limited. Mill for example, allows only this ‘one very simple principle’ that justifies placing limits on free speech: ‘the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others’ (1978, 9). There seems good reason to think that not only do public figures have a duty not to spread misinformation about COVID-19, but that it may also be justified to silence those who violate this duty in certain circumstances. During the pandemic, some social media companies did this. On 27 July 2020, for instance, Twitter deleted Trump’s retweet of a video titled ‘COVID has a cure. America wake up!’ This video was hosted by a group calling itself ‘America’s Frontline Doctors’, and advocated not wearing masks and for hydroxychloroquine as ‘a cure for COVID’.¹⁵

Related to these obligations, we propose that celebrities also have a reparative duty to try to undo harm they have caused by violating either of the two duties above. This means that if a celebrity is responsible for the spread of misinformation then they have a duty to try and repair the damage they have done by publicly stating that their previous claims were false or without evidence. For example, when British television presenter Eamonn Holmes appeared to express support for the conspiracy theory that 5G networks are responsible for the spread of coronavirus, many demanded that he publicly deny such a link (Manavis 2020). Similarly, after criticizing Oprah for directing trust towards ‘voices of dubious characters’ (cited above), MacFarlane then called on her to ‘lend her own powerful voice to correct it’ (Wilstein 2020). The duty here is a duty of reparation, a special duty that arises from the performance of a wrong act in the past (Ross 1930, 21). As Radzik (2009) argues, a key part of making amends for wrongdoing is making efforts to repair the damage that one has done. In this context, that means making an effort to correct mistaken beliefs and misdirected trust caused by one’s testimony.

6. Positive Duties of Celebrities

We might think that the best way for celebrities to ensure they do not violate their negative duties would be to adopt a policy of prudent silence about issues they do not understand. Particularly during a pandemic, the safest policy may appear to be for celebrities to keep quiet about how to respond and thereby avoid being blameworthy for violating these duties. However, we will argue that celebrities will also be subject to positive duties that result from their position of epistemic power.

In the original examples discussed by Fricker (2007, chapter 4), we saw that epistemic justice includes the disposition to pay special attention to the credence one lends to the testimonies of historically disadvantaged groups. This can enable us to correct our identity prejudices and so help to avoid committing a testimonial injustice. To put this in terms of duty, it seems reasonable to think that hearers have a negative duty to avoid committing testimonial injustice, and through this process of correction they can help to comply with this negative duty. However, testimonial justice may also generate positive duties. When someone witnesses another person committing a testimonial injustice they have good reason to try to intervene to help them correct against this. For example, if I hear a colleague dismiss a woman’s testimony and I suspect that this is a result of a sexist prejudice, then I may have a duty to try and help my colleague to correct against their prejudice.

Similarly, those who receive a larger amount of attention on a particular topic than they are due, may have a positive duty to try and redirect that attention to those who deserve it. When the testimonial injustice concerns the way in which attention is distributed, the way to correct this is to try and redistribute this attention to those who deserve it. Suppose a celebrity is asked about their opinion on coronavirus, as Klopp was in our earlier example. They could refuse to answer the question to ensure they do not violate any negative duties. But they could also use this opportunity to try to correct the unjust allocation of attention on this topic by redirecting attention to the appropriate experts.

We propose that those who receive undeserved attention on a particular topic have a general imperfect duty to seek to redirect attention towards those who deserve it. It is imperfect because people should adopt correcting for these kinds of testimonial injustices as a general end but do not have a duty to redistribute attention this way in every situation where the opportunity arises. However, during a pandemic there is good reason to think that celebrities may be subject to a perfect duty to redirect attention to those who deserve it, especially when asked to comment on the pandemic. In this situation, redistributing attention is not only a matter of testimonial justice but also minimizing the harms resulting from the pandemic. Here, celebrities have a duty to try to redirect the attention they receive to the proper experts.

In addition, celebrities may also have a positive imperfect duty to use the attention they receive to promote behavior that will reduce the spread of a pandemic. There are several ways they might fulfill this duty. First, celebrities may choose to take on the position of a role model. Think back to Schwarzenegger’s Instagram posts. Schwarzenegger used the attention he received to offer an instructive example of how to behave under pandemic conditions. Schwarzenegger can serve as a positive example which may encourage his fans to behave in ways that reduce the spread of the virus.

Moreover, Schwarzenegger’s posts allow him to act as a salience magnet. The effect of Schwarzenegger’s actions are magnified beyond one-to-one influence because each of his fans is in a position to know that his other fans may also be influenced by his posts, which would activate conditional rule-following as described above. When it comes to marshaling a response to a pandemic, norms such as self-isolation or vaccination are primarily worth following when they are widely adopted. Under these conditions, it is imperative for conditional norm-followers to receive reassurance both that (1) enough others are complying, and (2) enough others expect them to comply.

Celebrities can operate as salience magnets by coordinating large populations. Their effectiveness in this role increases in line with how believable their testimony is, as well as how quickly it can be shared (such as by digital technologies). This can work in two ways. Typically, the celebrity makes an assertion or directive concerning the best way to promote public health under pandemic conditions. This means that everyone in the celebrity’s audience is in a position to know the content of the assertion, and to know that everyone else is in a position to know it too. Nevertheless, an individual may then ask themselves: ‘Why should I believe the testimony of this celebrity?’ Unless the celebrity happens also to be a domain expert (such as an epidemiologist or public health expert like Anthony Fauci), it is difficult to find reasons for answering this question in the affirmative. Likewise, we might ask: ‘Why should anyone else believe or obey this celebrity?’ Again, the answer may well be: ‘No good reason’. If this is right, then celebrities who make assertions and issue directives are ill-suited to play the required role as cooperation-fostering salience magnets. This should remind us of Klopp’s insistence that he has no expertise to offer, and so should not be consulted for his opinion on the pandemic.

Nevertheless, we should resist giving up too quickly on the role of celebrities in these sorts of contexts. The medieval town church bells mentioned above do not possess expertise, but they still function as appropriate salience magnets. The call to prayer lacks expertise, but it also serves effectively as a salience magnet. Might the same be the case in the case of celebrities? To understand this idea, we must consider the second way in which a celebrity’s epistemic power might be harnessed to foster coordination and cooperation at scale. Celebrity power could be expressed not via linguistically mediated assertions and directives, but via behavioral modeling. In other words, instead of a public service announcement in which a celebrity conveys information about the pandemic or tells people what they ought to do, celebrities can visibly enact behaviors that public health experts deem necessary for population-level coordination in the face of the crisis. Celebrities are well suited to this role given the levels of attention they receive and because the amount of expertise required to model this behavior is much less than that required to issue authoritative advice.

Viewing celebrities as salience magnets differs from viewing them as straightforward role models. What makes someone a role model is their individual, intrinsic virtues, whereas to be a salience magnet one does not need any special virtue whatsoever. What is important is the position the individual occupies; that is, what matters is that they are famous. Likewise, the epistemic power argument starts from the fact that celebrities have power, and then asks how they should (or should not) wield it. By contrast, the salience magnet argument starts from the fact that in order to foster coordination and cooperation, groups need bases for common knowledge, and that highly salient objects or famous people can fulfil that requirement effectively. We might think of this role analogously to ones played in exercise classes, informal dance circles, and the children’s game of monkey see, monkey do.¹⁶ In many exercise classes, the group leader models the postures and movements that the rest of the class are required to imitate. To do this, the leader needs to know which movements to perform and how to do them, but they need not understand the physiological explanation for why these are the right movements. Indeed, they might, and probably often do, have wildly false beliefs about the justification involved. Likewise, when a group of people forms a dance circle and one of them moves to the center, the others might imitate their movements. It does not much matter to anyone what the movements are, so long as everyone is doing the same thing. Having a single individual serve as an exemplar works well in such a scenario, especially when they are a salience magnet in virtue of the position they occupy in the social structure (being situated at the center of the circle, for example).

One way to think of the use of celebrity epistemic power during a pandemic, or any other public health crisis in which many individuals have to cooperate, is by analogy to the role of the exercise instructor: there is a limited range of acceptable behavior for them to model, but it does not much matter whether they have (1) correct beliefs, (2) incorrect beliefs, or (3) no beliefs about how to justify correct behavior. We might draw a scope distinction and say that what is needed is that someone serves as a salience magnet, even though it’s not the case that there is someone who needs to serve as a salience magnet. Celebrities just so happen to be ready-to-hand and fit-for-purpose, not because of any special intrinsic qualities they embody, but just because what is needed from a structural point of view is someone who is likely to garner everyone’s attention and admiration.

Acting as a role model or a salience magnet are two ways of fulfilling an imperfect duty to promote behavior that will reduce the spread of a pandemic. Not all celebrities need to post self-quarantine clips on TikTok. Instead, what is needed is effective saturation: the prosocial actions of a few highly prominent celebrities will be enough to have a marked effect on public health under pandemic conditions if they can capture the attention of most in the population that needs to cooperate. Those who choose not to act as salience magnets themselves, may still make a positive contribution by directing attention towards others who are playing this role. A celebrity with no interest in posting self-quarantining videos may direct their followers’ attention to the videos of others.

7. Conclusion

We have considered how distributions of attention may impact upon ones testimonial duties by focusing on the case of celebrities during pandemic conditions. Celebrities can use or abuse the power derived from the attention paid to them. We argued that there are several negative duties, including the duty not to spread misinformation and the duty not to commit testimonial injustice by undermining trust in authoritative sources (or direct trust towards quacks or conmen). Stemming from these duties is a duty of reparation, which appears when a celebrity violates one of these negative duties. We also argued that celebrities have positive duties. First, they have a general imperfect duty to try to correct for an unjust distribution of attention by redirecting it to those who deserve it. In the case of a pandemic there seems good reason to think that this duty may become a perfect one, due to the harm that could be prevented if people follow the advice of experts. In addition, celebrities have an imperfect duty to promote behavior that will reduce the spread of a pandemic. We outlined three ways they might do so: they might take on the position of a role model, they may act as a salience magnet, or they can direct people’s attention towards others who have taken on these roles.

While our focus has been on celebrities, it is worth noting that some of these arguments also apply to others. The moral reasons that speak in favor of celebrities being valuable role models and using their epistemic power in a way that is in line with testimonial justice applies to everyone who has influence over how other people behave and what they believe. These reasons will be weaker for those with less influence and public attention. This means that those who have no special place in the public eye are less likely to have a major negative impact if they act against these moral reasons. The difference here, however, is a difference in degree rather than one of difference in kind. This means that those who are not in the public eye also have a responsibility to be careful about spreading misinformation and an imperfect duty to redistribute attention to those who deserve it and to model appropriate behavior during a pandemic.¹⁷

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The work was supported by the Alexander von Humboldt Foundation Australian Research Council [DP190101507]; John Templeton Foundation [61378].

Notes

1. Though there has been some discussion of the impact of this phenomenon on democracy (Archer et al. 2020), for the ethics of honoring and admiring (Archer and Matheson 2021), and for attentional excesses and deficits (Alfano and Skorburg 2018; Gardiner 2022; Smith and Archer 2020).

2. For those of the Bayesian inclination, this can be modeled as a real number between 0 and 1 inclusive. If I trust you completely (if I have credence 1 in you as a source), then when you assert that p I update my credence in p to 100%. By contrast, if I put no trust in you (if I have credence 0.5 in you as a source), then when you assert that p I don’t update my credence in p at all. Credences below 0.5 are theoretically possible though probably rare; they can be thought of as trusting someone to lie or otherwise speak falsely. In this framework, someone embodies a credibility deficit towards a source when the source deserves a credence of, say, 0.9 but enjoys only a credence of 0.7. Likewise, someone embodies a credibility excess towards a source when the source deserves a credence of, say, 0.6 but enjoys an inflated credence of 0.9. In other words, credibility deficits and excesses relate to whether people are disposed, once they receive an assertion, to trust the speaker’s word less than or more than the speaker deserves. And such deficits and excesses may constitute testimonial injustices when they are motivated or explained by unjust identity-based prejudices.

3. For background on the ethics of disaster, see Zack (2009).

4. Prominent individuals can also have duties derived from their roles. For example, New Zealand’s health minister, David Clark, was forced to apologize after being caught breaking quarantine to go mountain biking.

5. See Archer (2021) for a discussion of ethical issues that arise from this kind of identification.

6. Such testimony has recently been characterized as ‘epistemic trespassing’ (Ballantyne 2019) – a problem that may not be entirely novel but which is increasingly frequent and worrisome as traditional gatekeepers such as peer review are eroded by venues such as Medium, Substack, and pre-print archives.

7. Fricker (2007, 18) discusses credibility excesses as possible forms of testimonial injustice but does not think they will be especially widespread or troubling. Medina (2011) argues that credibility excesses are an important form of epistemic injustice, as they bestow people with undeserved epistemic privilege, and because one person’s excess is often directly pitted against another’s deficit.

8. Wellman’s argument is targeted at sporting celebrities but his point applies to celebrities generally.

9. See Schindler et al. (2013) for a helpful overview. One study conducted by van de Ven, Zeelenberg, and Pieters (2011) casts doubt on this connection. This study, however, appears to be something of an outlier. Indeed, van de Ven, the study’s lead author, has recently rejected his earlier conclusion (van de Ven 2017; van de Ven, Archer, and Engelen 2019).

10. This is a distinct notion from that of epistemic authority. An epistemic authority has been claimed to be someone who does what I would do if I were more conscientious and better at getting to the truth (Zagzebski 2012, 109). Alternatively, we might view an epistemic authority as someone who has novice-orientated abilities (Croce 2018). Either way, these views of epistemic authority view this authority as a praiseworthy attribute while ‘epistemic power’ is intended as a value-neutral term. The fact that someone possesses this power does not mean that they should possess it.

11. For more on this, see Lewis (1969), Dezecache and Dunbar (2012), Dunbar (2012), and Chwe (2001).

12. Wellman (2003, 333–334) considers the objection that special obligations can only be acquired voluntarily, but he points out that there does not seem to be good reason to accept this claim, as special obligations can arise, for example, simply from being in a uniquely good position to offer assistance.

13. Spurgin supports this point by asking us to imagine that children break into a house and witness two consenting adults having sex, leading the children to try and imitate this behavior. Spurgin (2012,120) claims that it would be unreasonable to blame the couple for their behavior. Likewise, Feezell (2005, 25) claims that the mere fact that someone is being imitated does not provide them with role model obligations.

14. This point is defended in more detail by Nagel (1998).

15. Smith and Niker (2021) argue that the fact that epistemically powerful agents are able to extend the reach of their voices through social media means social media companies have a regulatory role to play to limit the harm such agents can cause.

16. In this game, one child is ‘it’, and all the other children are meant to imitate their movements. It does not matter who is ‘it’, so long as everyone knows who is ‘it’. It does not much matter what the child who is ‘it’ does, so long as all the other children can safely imitate. It takes no special virtue or expertise to qualify as the one who is ‘it’.

17. Thanks to Colin Klein for helpful comments on an earlier draft of this paper. Alfred Archer’s research was funded by a grant from the Alexander von Humboldt Foundation. Mark Alfano’s research was funded by grants from the Australian Research Council and the John Templeton Foundation.