Weaponized Testimonial Injustice

. Theoretical tools aimed at making explicit the injustices suffered by certain socially disadvantaged groups might end up serving purposes which were not foreseen when the tools were first introduced. Nothing is inherently wrong with a shift in the scope of a theoretical tool: the popularization of a concept opens up the possibility of its use for several strategic purposes. The thesis that we defend in this paper is that some public figures cultivate a public persona for whom the conditions of the notion of testimonial injustice might be taken to apply, and this situation is exploited to their advantage, as a means to advance their political agendas. More specifically, they take advantage of this to generate situations of crossed disagreements, which in turn foster polarization.

who hired him had defended him previously when Scruton made similar statements. In particular, he was defended by James Brokenshire as a "champion of freedom of speech" and praised as someone who is often expressing strong and controversial views. Scruton perceived himself as a victim of our times, or that is suggested by his complaint that "thought crimes are being manufactured to hound conservatives out of public life" (Williams, 2019). In a similar way, Mario Vargas Llosa recently complained that the particular sociolect of certain minorities was silencing the Spanish language, victim of a campaign promoted by these minorities (Vargas Llosa, 2020). In all these cases, the scope of some theoretical tools aimed at making explicit the injustices suffered by certain socially disadvantaged groups ends up allegedly including situations in which the victims belong to privileged groups.
Nothing is inherently wrong with a shift in the scope of a theoretical tool. The popularization of a concept opens up the possibility of its use for several strategic purposes. There is no need to pass judgment on the legitimacy of these strategies to realize that the use mentioned above subverts the intention of those who started using the concept in the first place -to fight against certain injustices suffered by the disenfranchised. If we want our philosophical concepts to effectively contribute to a meliorative project of any kind, these concepts need to be adapted to an ever changing context. Sometimes this need shows a deficiency in the original characterization of the concept, but this is not necessarily the case. Some other times, bringing these unintended uses to the fore is enough in order to counter habituation, or a changing public opinion. Such is the purpose of this paper, to provide an exposé of certain communicative strategies, aided by the notion of testimonial injustice -or an intuitive understanding of it, that subvert the original intent of those who propose the notion in the first place, making it an accessory to a new form of injustice.
The concept that we propose to reflect on is that of testimonial injustice. We will focus on cases where a lack of credibility -crucially, not any kind of lack of credibility, but one that is perceived to be unfair and that makes them closer to a target audience, which can be plausibly construed by some as a case of testimonial injustice-is used to the advantage of political figures in a privileged position. Public figures that seemingly cultivate an unreliable persona, able to say anything at any moment, someone from whom anything can be expected -and exploit this public persona to their advantage, as a tool to advance their political agendas. Explicit appropriation of the concept of testimonial injustice is by no means the main feature of the cases that we pay attention to in this paper. There might be explicit appropriation, to a certain degree, but only incidentally -the goal is to take advantage of the lack of credibility to be able to subvert a specific kind of public debates, highly monitored dialectic scenarios which are supposed to flesh out deliberative democracy, turned instead into perfect opportunities to promote polarization. Explicit appropriation of the concept of testimonial injustice, to present these figures as victims, sugar-coats the strategy. An explicit or implicit use of the notion of testimonial injustice is, nevertheless, essential to our description of the phenomenon. Alleged victims are not perceived as looneys, they use their well-earned lack of credibility to generate a cohesive, identity-fueled following, together with a permanent out-of-jail card to turn difficult debates into opportunities to advance their agendas.
The thesis that we defend in this paper is that some public figures cultivate a public persona for whom the conditions of the notion of testimonial injustice arguably apply, and that this is subsequently used to take advantage of the constructed image of someone unreliable, someone from whom everything can be expected, to blur the debates in which they appear. The benefits of the constructed image for which the conditions of the notion of testimonial injustice arguably apply, we will argue, are achieved via generating situations of crossed disagreements. This strategic use of the concept of testimonial injustice is not necessarily a calculated maneuver. The public persona created for this purpose does not necessarily reflect on the real personality of those performing this role -the properties of the public figure do not automatically translate into character traits of the person, nor the other way around.
To be clear, we neither argue that the concept of testimonial injustice overgeneralizes nor that it needs to be amended. We just point out that the concept is put to use towards an aim that differs from the one it was originally intended to, and that we need to take this into consideration. Imagine a tennis racket being used to hit somebody in the head, or a Montblanc fountain pen used to stir a cocktail. The object can certainly be used to perform an activity for which it was not intended to, and it might be terribly effective at this new task, even if this activity goes somehow against the original aim for which it was designed. None of this means that there is something wrong with the original design of the object. This, we claim, is the case with testimonial injustice in our paper: the strategy that we identify only works because the concept of testimonial injustice is used beyond its original aim. The examples provided in this paper are just meant to display how the mechanism works. We take this to be a much more general phenomenon, one that is not restricted to the few examples that we mention here. Of course, our claim is not that this feature exhausts the complex behaviour of the kind of political figures that we will mention. It is neither our goal to show that the strategy that we will explore in this paper explains on its own the success of these figures, or their general approach to politics. It's just something that they do, something that serves them particularly well.
In section A, we introduce the notion of testimonial injustice, and the phenomenon of crossed disagreement. In particular, we review the conditions that a case of credibility deficit must meet in order to count as a case of testimonial injustice, and explore how the phenomenon of crossed disagreements conflicts with the liberal ideal of political knowledge. In section B, we present some examples of how some political figures cultivate a fringe political persona, and discuss how the resulting image meets the conditions of the notion of testimonial injustice. Section C presents some specific examples in which being perceived as someone unreliable, someone from whom certain outbursts can be expected, is used to disrupt a public discussion, and prevent the other party from conveying the information that she was trying to communicate.

Testimonial injustices
"Testimonial injustice" is the well-known label introduced by Miranda Fricker to refer to a very specific unjust situation, the injustice of receiving, in a systematic and persistent way, less credibility than one deserves (Fricker, 2007). According to Fricker, receiving a reduced level of credibility when making a claim counts as a case of testimonial injustice just in case the speaker's reliability is diminished because of the hearer's prejudices about the speaker's social identity, be it in terms of her gender, race, sexual preference, etc. (ib., p. 17). It is the speaker's belonging to a particular social group which leads the hearer to attribute her less competence or sincerity, and therefore a lower level of credibility (ib., p. 45). In other words, it is the identity prejudice that causes the credibility deficit, and establishes the systematicity and persistence conditions: identity prejudices affect different dimensions of social life and make the credibility deficit persistent (ib., p. 27).
Testimonial injustice is a type of epistemic injustice because the harm perpetrated is epistemic in nature: the speaker is damaged as a knower, specifically in her capacity as a giver of knowledge (ib., p. 44) 4 . One of Fricker's most famous examples of testimonial injustice is the case of Tom Robinson in Harper Lee's To Kill a Mockingbird. Robinson, who is an African-American, is prosecuted for allegedly raping a white woman. Despite the fact that the evidence available supports Robinson's innocence, or is at best far from being conclusive concerning his involvement in the rape, the jury, guided by their prejudices against African-Americans, takes him to be guilty of the charge (Fricker, 2007, p. 23). Epistemic injustice can take place in different ways and to different degrees, ranging from the blatant to the most subtle, and it can at times be difficult to identify. Every case of epistemic injustice exhibits certain common features, though, such as underestimating someone's testimonies and interpretations because of the group that the speaker belongs to. This might affect radically different groups. Credibility deficit suffered by ill people, for example, can count as a case of testimonial injustice too (Carel & Kidd, 2014;Crichton et al., 2017). For instance, patients of myalgic encephalomyelitis/chronic fatigue syndrome usually receive less credibility than they deserve simply because they belong to a specific group of ill people (Carel & Kidd, 2014, p. 4).
The theoretical landscape covering different but related forms of epistemic injustice has been expanded considerably in the past few years. New forms of injustice have been identified, potential forms of epistemic injustice that go beyond testimonial (and hermeneutical) ones. Dotson introduced a different form of epistemic injustice, called contributory injustice (Dotson, 2012), which occurs when a person has the conceptual tools to understand and communicate her experience of oppression, but her testimony is undermined because her audience willfully misunderstands her. Another different form of epistemic injustice, called conceptual competence injustice, occurs when an epistemic agent makes a conceptual claim and is wrongly perceived as lacking the appropriate conceptual or linguistic competence (Anderson, 2017). There are also epistemic injustices focused not on the identity of the speaker, but rather on the things communicated by the speaker, named content-focused injustice (Dembroff & Whitcomb, forthcoming). Another different phenomenon is epistemic appropriation (Davis, 2018). Epistemic appropriation comprises two epistemic harms: epistemic detachment and epistemic misdirection. The former occurs when epistemic resources developed within the margins are overtly detached from the marginalized knowers responsible for their production as they gain intercommunal uptake. The latter happens when epistemic resources developed within, but detached from, the margins are utilized in dominant discourses in ways that disproportionately benefit the powerful. Together, both harms constitute epistemic appropriation. Finally, another more subtle form of epistemic injustice -situations where victims of oppression are asked to let go of the anger and other feelings they experience in response to the injustices they face in order to improve their chances of putting an end to the situations they encounter-have been pointed out as affective injustice situations (Srinivasan, 2017;Archer & Mills, 2020).
Testimonial injustice, as we saw, has under its scope situations in which agents' words unjustly receive less credibility than they deserved, or the agent's epistemic capacity is undermined, due to the identity prejudices of the hearer. A closely related but slightly different injustice occurs when the speaker cannot perform some actions with her words simply because of her social identity (Kukla, 2014), and as a consequence her range of available speech acts is unjustly limited because of the unjust norms governing our social practices (Ayala, 2016). Injustice in this case does not concern the agent as a knower -it does not affect her ability to give testimony-, but as a speaker -it has to do with her ability to perform a given speech act. If, for instance, a female judge tries to denounce the banner used by the Spanish Supreme Court to announce an event as sexist and discriminatory and her words do not receive the appropriate uptake, in spite of the fact that her claim fits a standard formula, and instead is taken as an expression of her feelings (Ayala, 2018), then the female judge has been a victim of a discursive injustice (Kukla, 2014). Notice that the extension of the concepts of testimonial injustice and discursive injustice are not the same. However, both depend on restrictions imposed to the node that a speaker occupies in the social structure, which configures and limits the meaning that she can convey through her words (Ayala, 2016(Ayala, , 2018Kukla, 2014). These restrictions apply systematically, are linked to the identity associated with the node, and chase the agents over other spheres of their activity, besides a particular context. In this sense, the notion of discursive injustice is an extension of the notion of testimonial injustice.
We focus here on the concept of testimonial injustice, and our analysis is not aimed at justifying a revision of the notion. Our goal is to show that the status of certain agents can be interpreted as the result of testimonial injustice while, at the same time, providing an opportunity to surreptitiously advance a particular political agenda. This neither shows any shortcomings in the way the notion is currently characterized, nor describes a new kind of injustice. It's, in this sense, different from the cases that we mentioned above. In the examples that we will explore, the notion does not overgeneralize, it does not prove to have an extension larger than it was supposed to, nor is it "appropriated". The phenomenon that we focus on showcases individuals turning around testimonial injustice to their own advantage. Not just any kind of lack of credibility, but the kind of testimonial deficit that might arguably happen to an individual because they belong to a certain disenfranchised group, for the kind of reasons that might be felt to chase them in other spheres of life, as members of that group. Only when this lack of credibility is perceived as an instance of how a group is systematically mistreated, can these individuals use it to their advantage, transforming potentially harming communicative situations into crossed disagreements that advance their political goals.
In order to introduce our cases, we need, first, to show that the credibility deficit that these agents endure can be reasonably interpreted as a case of testimonial injustice. By this all we mean is that those who take it to be a case of testimonial injustice, especially those who belong to the group that is supposed to be on the losing side of things, have their reasons, their own assumptions, and therefore similar behavior can -and perhaps should-be expected in similar situations. Again, we do not think that they are right -some crucial assumptions are false, to our lights, but that there is nothing irrational in thinking along those lines. As we have pointed out, and as Fricker herself emphasizes (Fricker, 2007, pp. 21-22), not every situation of credibility deficit is a testimonial injustice, even if the credibility deficit is unjust and harmful. There are at least three requirements that a situation of credibility deficit must meet in order to be considered an instance of testimonial injustice, instead of epistemic bad luck: 1. Pertaining: the victim must pertain to a particular social identity or group that locates them in a specific node of the social structure in which their capacity to give knowledge, to be an informant or to convey certain meaning is undermined. 2. Systematicity: the social identity or group to which the victim belongs must be subjected to an unjust and systematic treatment through different social dimensions of social activity (educational, religious, sexual, political, legal, professional, economic, etc.). 3. Morality: the harm caused to the victim by the wrongdoers must be ethically culpable, i.e., it must constitute a violation of some moral principle (Ayala, 2016, p. 883;Kukla, 2014, p. 455;Fricker, 2007, pp. 21-22).
It is the purpose of this paper to show how social awareness around cases of epistemic injustice, unfair attributions of credibility to a given agent, might be used for the benefit of those in power. More specifically, presenting oneself as a victim of epistemic injustice might be the key to promote a particular political agenda, while seemingly contributing at the same time to the development of certain liberal ideals.

The liberal ideal of political knowledge
How should decisions that affect the common good be adopted within contemporary liberal democracies? A democratic society is a community where different voices, and not only some privileged ones, must share power in setting their common fate. A democratic system involves the participation of the citizenry at some fundamental level in the direct or indirect selection of laws, governments and the like, and deliberative democracy is assumed to be one of the best expressions of such a system. Decisions that affect the group must be reached through public deliberation, in an institutional setting that provides checks and balances; different opinions and forms of life must be taken into account in public deliberation processes. Knowing what to do under such a democratic ruling requires inclusive discussions where ideally everyone has a saying. One of the classical sources that support this ideal is Mill's conception of knowledge. According to Mill, knowledge requires being able to take all perspectives into consideration: "hearing what can be said about it by persons of every variety of opinion, and studying all modes in which it can be looked at by every character of mind" (Mill, 2009, p. 27). Political knowledge is no exception.
Within this liberal ideal of political knowledge, disagreeing is particularly valuable -it is only via disagreeing with those who oppose our views that we get to actually know. In order to achieve knowledge, rather than mere belief, it is necessary to be engaged in the public game of giving and asking for reasons. We need confrontation with those who disagree with us to be able to distinguish between what seems correct to us and what is correct (cfr. Wittgenstein PI § §185-202). As Medina puts it: In order to properly elucidate our normative activities and to provide normative guidance for our interactions, we should avoid idealizations and go back to the rough ground of our actual practices where we find differently situated knowledges and perspectives-where there is friction. (Medina, 2013, p. 11) Correspondingly, only those decisions that are adopted as a result of an inclusive process of discussion will be taken to contribute to the democratic ideal of political deliberation. Contemporary liberal democracies aim at creating a stable social and political order where disagreeing plays an essential role, and every opinion is worth consideration. Thus, from this liberal point of view, silencing a voice in public deliberation is epistemically and politically detrimental. In Mill's words: "If the opinion is right, they are deprived of the opportunity of exchanging error for truth: if wrong, they lose, what is almost as great a benefit, the clearer perception and livelier impression of truth, produced by its collision with error" (Mill, 2009, p. 30).
We will argue that, under certain circumstances, the explicit goal of integrating everyone's voice in a public deliberation that follows from this liberal notion of political knowledge can lead to situations that undermine the well-functioning of democracy. It is important to note, however, that our point is not that the liberal notion of political knowledge is inconsistent. Rather, in the rest of section 2 we will only claim that this ideal, in particular contexts, might affect the functioning of democratic institutions in undesired unexpected ways, boosting the practical challenges that a polarized public opinion poses to the normal functioning of certain democratic institutions, such as increasing the difficulty to coordinate and cooperate, and diminishing societal trust (Carothers & O'Donohue, 2019; Levitsky & Ziblatt, 2018;McCoy & Somer, 2019).

Disagreement and polarization
At times, being exposed to those who oppose our views gets us further away from the other part, especially when the body of evidence we encounter is mixed (Gilovich, 1991, p. 55;Kelly, 2008;Lord et al. 1979). A recent empirical study (Bail et al., 2018), to mention one example, showed that being exposed, through Twitter, to the opinions of someone who is sympathetic to an opposed political ideology makes us more polarized, rather than less. In the study, participants (Democrats and Republicans) were asked to follow on Twitter, during one month, a bot that would retweet a total of 24 messages per month, randomly sampled from a list of political Twitter accounts (e.g., opinion leaders, media organizations, etc.). The Twitter bot that Democrats would follow was a "conservative" one -it only retweeted messages from accounts with conservative content. The account that Republicans would follow was a "liberal" account -it retweeted messages from Twitter accounts that shared just liberal content. At the end of the experiment, they found that Republican participants became substantially more conservative, and Democrats also showed a slight increase in their liberal attitudes.
Diversity of opinion does not automatically translate into a meaningful exchange of views, one that will take us closer to reaching a consensus. Coordination and knowledge in a democratic environment require more from deliberation than sheer heterogeneity in the composition of the group of interlocutors. As the above study suggests, a person might become polarized when she is exposed to the opinions of those with different political ideologies, instead of moving her way towards a common view. In certain situations, especially when our interlocutor is perceived as radically different from us, exposure to opposing views can contribute to polarization (Bail et al., 2018;Wallace, 2016;Yardi & Boyd, 2010). We can become more distanced from others when we get exposed to the arguments of those who do not think like us, and evidence challenging our beliefs, because reacting to information from "the other side" might reinforce some of the arguments supporting our position (Yardi & Boyd, 2010), and we may even add new arguments to the set of arguments that support our stance. Disagreement, thus, doesn't necessarily bring us closer to reaching a consensus towards a common course of action in a democratic setting. In fact, there are situations where the very nature of the disagreement may carry politically undesirable consequences, such as an increased radicalization of, on the one hand, the subjects involved in the dispute and, on the other, those that witness that controversy.
In a public disagreement, the parties involved in the dispute usually rehearse different arguments to provide support for their position, and these arguments often belong to different categories: sometimes they are about matters of fact, other times they are based on normative or value issues, etc. Depending on the object of discussion and the nature of the information at stake, the debate might contain a mixed bag of different types of disagreement -factual if it is about whether some states of affairs are the case, metalinguistic if it is about the meaning of a particular term, deep if it is about the standards, etc. We follow here a simplified taxonomy of disagreements proposed by Osorio and Villanueva, which focuses on the way a disagreement might be conceived. They distinguish three types of disagreement: Type A. Disagreements in which there is a presumption of commonality with respect to the standards of both parts. Roughly, fact-dependent disagreements.
Type B. Disagreements that become about the standards, once it becomes obvious that both parties have different standards. Roughly, deep disagreements.
Type C. Disagreements that neither disappear nor become about the standards, once it becomes obvious that both parties have different standards. Evaluative disagreements. (Osorio & Villanueva 2019 118) Normally, the topic of disagreement is studied in static terms, i.e., by analyzing the nature of a discussion at a particular time. However, our natural disputes have a dynamic -and heterogeneous-character, and the subjects involved in them do not actually register and react to every piece of information provided, because it is really hard to do that in real-life linguistic exchanges. They might, consequently, end up exhibiting clear signs of conceiving the dispute differently, even if they are not aware of that. When this happens, the parties involved in a public dispute are engaged in a crossed disagreement (Osorio & Villanueva, 2019, p. 120).
Crossed disagreements are not instances of talking at cross-purposes; they are genuine disputes, where the parties in contention conceive the disagreement in significantly different terms. It will, for instance, count as a crossed disagreement when one party thinks of the discussion in factual terms, as if it is a disagreement of type A, and the other conceives it as a normative dispute, i.e., as a type-B disagreement. As pointed above, most interventions in public disagreements showcase a plurality of arguments, and it is therefore only to be expected that these are heterogeneous in nature -some factual, some normative, etc. This, however, does not prevent a disagreement from being crossed in nature, when the set offered by one of the parties is not properly matched by the response of their interlocutors.
When the interlocutors in a dispute exhibit signs of conceiving differently the nature of the disagreement they are involved in, as a result they do not actually take into account the arguments offered by the other part, which on its own might make it more difficult to reach a decision on a common course of action. Besides, crossed disagreements expand the pool of arguments supporting the views held by each party, either by making the set bigger or by repeating some of the arguments already contained in it (Osorio & Villanueva, 2019, p. 126). Increasing the size and density of the pool of arguments to which one is exposed fosters polarization: when we are exposed to a limited and biased 5 set of arguments, especially when they are frequently repeated, we tend to become more impervious to others' reasons (Sunstein, 2017;Barberá et al., 2015;Vicario et al,. 2016). Of course, this argument pool does not increase with arguments for and against your position, the key point is that it grows asymmetrically: it increases with respect to the arguments in favor of our initial view, but not to the arguments against it. The resulting set of arguments is partial in this sense. Crossed disagreements are instances of disputes that covertly impede real coordination, because they appear to contribute to the deliberative process -therefore seemingly making the chances of reaching a consensus higher-, while furthering at the same time the distance between those who share the core tenets of the parties involved in the dispute.
According to the studies conducted by the Centro de Investigaciones Sociológicas (CIS), polarization in Spain increased between 2008 and 2012 regarding the territorial organizaton of the Spanish state prefered by the citizens (Barómetro, 2020). The support for the option "A state with a single central government, without autonomy" went, on the one hand, from 8.6% in 2007 to 24.9% in 2012. The support for the opposite option on the ideological spectrum, "The recognition as independent states by the autonomous communities" went, on the other hand, from 17.4% in 2008 to 28.2% in 2012. During this period, the ratio of crossed disagreements that took place in the Parliament systematically rose, and it was higher than before 2008 and after 2012. There was a high correlation between the presence of crossed disagreements in the Parliament and the increase in polarization in Spain with respect to the territorial issue (Almagro et al., ms.). This we take to be evidence of the connection, both empirical and factual, between crossed disagreements and polarization.
The findings reviewed in this section help to partially explain why crossed disagreements foster polarization: we do not reject the information that comes from the other party, but the effort to reinterpret and criticize it, especially in a crossed-disagreement situation, makes us more polarized because it increases the size and density of the argument pool we are exposed to, and therefore reinforces our previously-held political preferences. To the extent that decisions that affect our common interest must be adopted by public deliberation, in an institutional setting that provides checks and balances, the interventions designed to ensure the proper development of public deliberation in a democracy must take into account the effects produced by crossed disagreements, which frequently go unnoticed in our public discussions.

B. When personality gets in the way
It is the goal of this paper to show that, under certain conditions, public awareness of the phenomenon of testimonial injustice can be taken advantage of by privileged parties in order to foster their political agendas. This is achieved, as we will see, by turning perfectly legitimate public debates into crossed disagreements, which in turn foster polarization. We now move to the kind of agents who are placed in the unique position to secure this, while seemingly contributing at the same time to the ideal of deliberative democracy. They are influential political agents able to promote a public persona that embraces the -sometimes extreme-disdain of the opposition, and feels for the most part no need to meticulously discredit those who accuse them of lying. A fact that is well known by their detractors, but also by their followers: they will say anything! Donald Trump and Boris Johnson are cited as examples of politicians who are prone to the epistemic vice of epistemic insouciance, "an indifference or lack of concern with respect to whether their claims are grounded in reality or the evidence" (Cassam, 2019, p. 79). Both political leaders, as we will show below, usually appear in the media making statements that are, to put it mildly, controversial and false. In fact, part of what characterizes them politically is their outbursts. The messages they convey in their speeches and the things they are able to do in the debates they are involved in are closely linked to their manners, and the media image they build 5 We are not using the term "bias" in the sense of the term that assimilates it to "irrationality", but we are using it as a synonym of "partial", and in line with several psychological experiments carried out about the interpretation and assessment of new evidence in such a way as to be assimilated into the preexisting assumption and expectations of a subject (Gerber & Green, 1999;Lord & Taylor, 2009;Taber et al., 2009). This particular evaluation of the evidence is the way in which "participants typically find evidence that is consistent with their prior beliefs more convincing than evidence that is inconsistent with their prior beliefs" (Jern et al., 2014, p. 218).
based on their unusual behavior. Crucially, their eccentricities are met with sympathy and promote their connection with those who believe that they share a common identity with these figures -they can appreciate how weird or off the mark some of their reactions seem to be, but take them to be, at least partially, a natural reaction to the kind of discrimination that they also seemingly suffer.
One might wonder why this well-known trait of the public personae they promote should affect political deliberation, particularly in an informational environment where reliable sources are not scarce. At first glance, one should expect well-informed people to treat reliable information sensibly. If a leader representing your political ideals provides information that proves to be false, the trust attributed to them should be undermined. In contrast, if a leader who does not represent your political ideals provides some true information, the confidence attributed to them should increase. But that does not seem to be the case. Empirical results on how people treat politically congruent, neutral and incongruent fake news suggest another explanation. In several experiments on the relation between fake news and political reputation, carried out by Altay and his colleagues (Altay et al., 2019), respondents treated politically congruent and politically neutral news in a similar manner, but not politically incongruent news. Participants did not lower their trust when they were confronted with politically congruent fake news. Moreover, participants failed to increase their trust when a politically incongruent real news was presented (Altay et al., 2019, p. 18). In a nutshell, recognizing the lack of precision in a leader does not automatically alter your confidence in them. Becoming aware of the lack of precision in some of the beliefs that support your political preferences does not necessarily lead to a revision of those political preferences. In addition, according to the findings of two recent studies (Porter et al., 2019; see also Nyhan et al., 2020), exposure to evidence that does not support our previous factual beliefs might actually change those beliefs to make them more factually accurate, without affecting our associated attitudes and/or policy preferences. For instance, in these studies respondents were presented with Trump's misstatements on climate change followed by a fact-check. The correction made the participants become more accurate, however, and surprisingly, their policy preferences regarding climate change remained unaffected. These findings allow us to qualify the phenomenon of the so-called "backfire effect" : even when we are open to correcting our factual beliefs, our policy preferences and associated attitudes are not necessarily affected. The information from the other party, when openly contradicts what we already think we know -as opposed to being somewhat ambiguous (see Gilovich, 1991, p. 49 and ff.), does not make us instantly more polarized, but it does not change our political preferences straightaway, either. Only when it is presented under certain guises, as in a crossed disagreement, new information is arguably linked to the rise of polarization. In the cases that we explore, lack of precision does not even reflect negatively on the perceived personality of the leader, who is taken to be a victim of the same kind of injustice allegedly experimented by their followers, inflicted by the elites, the media, etc., and who takes advantage of the situation to generate crossed disagreements.
As pointed out above, the Prime Minister of the United Kingdom, Boris Johnson, is one of the political leaders who has been associated with the epistemic vice of epistemic insouciance, exhibited through a sort of erratic public behaviour that might stretch from outright lying to openly inappropriate demeanor -always with a smile. For instance, in 2017, during a visit to the Shwedagon Pagoda, the most sacred Buddhist site in Yangon, Myanmar, Boris Johnson started reciting the opening verses to Rudyard Kipling's Mandalay, a poem that speaks of the nostalgia of a retired serviceman looking back to his colonial service. Due to the romanticizing nature of the poem about colonial exploits of the British Empire, the UK ambassador to Myanmar, Andre Patrick, tried to stop Boris Johnson repeatedly, claiming that "it is not appropriate", to no avail. In his early career as a journalist, to name another example, Johnson gained fame for writing articles that were often unrelated to reality. He was fired for lying more than one time. One of his fellow journalists, Mathew Parris, described him as "under-prepared, jolly, sly, dishonest and unapologetic" (Parris, 2016; see also Cassam, 2019, p. 79). Donald Trump is another politician who exemplifies this kind of lack of concern, a trait that has stirred more than one public controversy. From climate change denialism (he literally said that "the concept of global warming was created by and for the Chinese in order to make U.S. manufacturing non-competitive") to racism, sexism and misogyny, Trump's behavior is shocking, to a degree, for most of us. Moreover, he uses social media as an amplifier for his political claims. With respect to the Sars-Cov-2 crisis, Trump has used his social networks to minimize the impact of Covid-19, to blame the Chinese government for the pandemic (he used an offensive label to refer to it) and even to advocate for pseudotherapies to treat the disease.
Partly due to this public behavior, that might be deemed "erratic" by some, Trump's credibility has gradually declined. A recent study from KFF Health Tracking Poll (Hamel et al., 2020) found that less than 50% of people trust Trump to "provide reliable information on coronavirus"; and just 36% of the people surveyed considered Trump as "honest and trustworthy". Our claim is not that taking advantage of a lack of credibility is the only reason they exhibit this sort of behavior. Neither it is that cultivating this public persona in the fringe is always done intentionally. Harm is done, though, by taking advantage of this condition, exploiting the lack of credibility that results from their unreliable behavior, in a context in which this is perceived as the result of an injustice perpetrated towards one of your own, to defend and push forward a specific political agenda.
As we have pointed out, there are at least three requirements that a situation where a person receives less credibility than usual must fit in order to be plausibly conceived as a case of epistemic injustice: i) the victim must pertain to a particular social identity or group that locates her in a socio-normative node in which her capacity to give knowledge is undermined; ii) the social identity or the group to which the victim belongs must, in turn, be subjected to unfair treatment with respect to different social factors (religious, sexual, educational, etc.) and iii) the harm caused to the victim must be ethically reprehensible, that is, it must constitute a violation of some moral principle (Ayala, 2016, p. 883). One might think that the cases of subjects discussed above, involving the likes of Boris Johnson or Donald Trump, do not exhibit these necessary features to be considered cases of testimonial injustice. There might be good reasons to maintain that those features do not actually apply to these cases, but they and their followers might also have reasons to think that they do. After all, they feel that they are being targeted by certain minorities -liberal elites, or whatever have you, accompanied by a blind mob. They feel that they are in fact victims of some sort of "political correctness", or "cancel culture", that forbids them from freely expressing their thoughts, and makes them less able to provide testimony. In August 2015, Trump said during an appearance on "Meet the Press" that "this political correctness is just absolutely killing us as a country. You can't say anything. Anything you say today, they'll find a reason why it's not good" (Cillizza, 2018). Their assumptions are most likely wrong; but provided that they have them, it is not irrational to see themselves as victims of testimonial injustice. Again, this has nothing to do with the possible shortcomings of the notion of testimonial injustice, it only shows how that notion can be put to use, once you set the stage in a certain way.
"Reasonable", as we have pointed out, only means that, given certain assumptions, it is a possible pattern of inference that they are victims of testimonial injustice, and that the same can be expected from those with the same assumptions, as, for instance, their followers. Of course, the particular assumption playing a role might turn out to be unjustified, or false, but it does not make the belief irrational, or the inferential pattern invalid. Take the following example. I think that my sister is arriving to the airport at 17:00, so I make a series of arrangements in order to be able to be at the airport by that time, since it has been a long time since we saw each other and many things need to be sorted out for me to be there at that time. I clear my schedule before, and after 17:00, so that I have plenty of time to take appointments later on. Everything works fine and I manage to be at the airport at 17:00. Yet, I did not know that scheduled arrival time indicates landing time, and I only get to meet my sister at 17:30. That means that I am late for an important appointment, at 18:30. My planning was reasonable, given what I knew, but I ended up in a bad spot, workwise. I should have known that scheduled arrival time meant landing time, and plan accordingly, but does this wrong assumption make my whole planning unreasonable or my belief that I would be on time for the meeting irrational? No. What makes my planning reasonable is that it can be explained in terms of reasons, and that similar behavior can be expected from someone with the same assumptions and the same goals. But of course I can still be wrong, and I can be proven to be mistaken if I share my plans with someone who actually knows that arrival time means landing time. Similarly, the agent's supporters might make reasonable inferences out of what they think that they know about their situation, and they are reinforced in those beliefs -they are outcasts, as their leader, they are disenfranchised, they are marginalized by corrupt elite media etc.
Thus, once certain assumptions are granted, one does not have to be crazy to see themselves as a victim under these circumstances. Firstly, it would be possible to argue that they actually belong to a disenfranchised group -their capacity to give knowledge or to be an informant is undermined because they belong to the subgroup of the "brave, politically incorrect" people. Secondly, they think that the group to which they belong is subject to systematic marginalization, their social group is a direct victim of unfair treatment. Thirdly, they think that the harm caused to them is ethically reprehensible. They think that their moral status is being questioned because they are "silenced" in one way or another. In May 2020, Trump wrote a tweet where he denounced the alleged in-mail vote fraud in the U.S. elections. Twitter placed a label on the tweet, warning that such claims could be false. In response to this, Trump accused Twitter of attacking his "freedom of speech" and threatened the platform to take action to regulate the social network. Because of this, Trump signed an executive order to fight against "selective censorship that is harming our national discourse" (Preventing Online Censorship, 2020). When these complaints are made explicit, we encounter cases of epistemic misdirection (Davis, 2018), where the conceptual resources developed by the disenfranchised are abused by people in power. The problem that we identify with this erratic behavior is a different one. Cultivating a public persona on the fringe, one from whom anything can be expected at a given moment, plus one who is rightfully perceived as epistemically insouciance, provides an advantage to those willing to promote a particular political agenda.
These agents are able to take part in constricted public debates, seemingly contributing to the deliberative ideal, while at the same time focusing on pounding on the arguments that their group already respects, irrespectively of the circumstances of the debate. Anything can be expected from them, and they honor these expectations. It's only in this context that the agents' behavior gets to be interpreted as a case of testimonial injustice, and only because of this his actions get to have the political impact that they have, instead of being the mere rant of a person who is perceived to be off the rails. The agent's strategy is all the more persuasive because the followers also feel left aside. The agent is presenting himself as an outcast, a victim of the sort of phenomenon that we might deemed testimonial injustice, not just as a mere loony, and we think that this is precisely what makes his strategy successful. Compare Kanye West's "meltdown" with one of Trump's outbursts. Kanye's case, as opposed to the type of cases we point to in this paper, is precisely the kind of case where an audience explains the situation purely on individual terms (e.g., "he's lost it") and, crucially, nobody listens to what he so strongly thinks he has to say. The world is full of extravagant people that nobody listens to. What makes these cases particularly powerful is precisely that this supposed exclusion allows the further strategy to work. A weaponized notion of testimonial injustice is what allows these agents to turn eccentricity and insouciance into a successful -anti-democratic-political strategy.

C. Crossed disagreements in key public debates
Crossed disagreements are conversational situations where the parties involved are having genuine disagreements, but each party seems to be conceiving the dispute as being one of a different nature, e.g., one sees it as a disagreement that can be settled by appealing to facts, while the other understands it as a dispute that might continue even if total agreement on the facts were possible. In which sense a public persona with the features described above can readily turn even strictly managed debates into crossed disagreements? The best scenario to show this is via public debates within the realm of contemporary deliberative democracies, public debates where a contribution to the liberal ideal of political knowledge is expected. Public political discussions are seen as tools that a society has to debate those topics that are important for our social life; these deliberative tools are capable of generating political consensus, decisions to adopt joint courses of action that everyone accepts as the result of the deliberative process. Even when a consensus is not reached, they are supposed to foster a better understanding of the nuances of alternative positions.
If these public debates do not work properly, one of the practical consequences they can have is an increase in polarization. As we have seen, crossed disagreements increase the size and density of the argument pool that the parts in dispute are exposed to: supporters of one side of the disagreement do not really address the arguments and reasons offered by the other part because they are of a different nature, and therefore they are more exposed to the ideas that already support their previous beliefs. The more you are exposed to the arguments and reasons that make you believe that you're right, the more polarized you become. Thus, crossed disagreements reinforce the political preferences held by each party, and, therefore, facilitate a lack of coordination. In this sense, crossed disagreements are instances of discussions that might covertly impede real coordination, because they appear to contribute to the deliberative process, while furthering at the same time the distance between those who share the core tenets of the parties involved in the dispute. Let us look at some real-life examples of debates that turned into crossed disagreements.

Case 1
During the first 2016 presidential debate, Donald Trump and Hilary Clinton were discussing how to improve the economic prospects of the country. In particular, the first question at stake was: "Why are you a better choice than your opponent to create the kinds of jobs that will put more money into the pockets of American workers?" At first glance, this question does not strike as a factual one: two people can agree on all the relevant facts and still disagree on whether Trump is a better choice than Clinton to improve the country's economy, provided they don't share a common conception on how economic success is, or should be measured. So, a disagreement on this issue might count as a type-B disagreement: a disagreement about what principle we should follow, how we should conceive the economic goals of a government, and what indicators we should pay attention to.
At the outset, Clinton shows clear signs of conceiving the dispute as a non-factual one. For clarity sake, we will italicize non-factual statements and underline factual statements: CLINTON: The central question in this election is really what kind of country we want to be and what kind of future we'll build together.
Arguably, the question of what kind of country we want to be and what kind of future we will build together is not a factual, but a normative one. Even if we agree on how our country actually is, we might still disagree on how we want it to be, or how it should be. Clinton argues that the best way to improve the economy is by trying to achieve a fairer distribution of resources, one that works for everyone, not just for the rich. To achieve greater equality, big businesses need to contribute by paying higher taxes. Trump, on the other hand, argues that the best way to improve the economy is to lower taxes to big businesses from 35% to 15%, and encourage them to expand and start other businesses. After arguing for a while that companies are leaving the country and that the U.S. is in deep trouble, Trump begins to interrupt his interlocutor frequently, and to show clear signs that, contrary to Clinton, he is conceiving the disagreement in factual terms, as if it were a type-A disagreement. To see this, consider the following exchange: CLINTON: Well, that's your opinion. That is your opinion. TRUMP: You go to New England, you go to Ohio, Pennsylvania, you go anywhere you want, Secretary Clinton, and you will see devastation where manufacture is down 30, 40, sometimes 50 percent. NAFTA is the worst trade deal maybe ever signed anywhere, but certainly ever signed in this country. And now you want to approve Trans-Pacific Partnership. You were totally in favor of it. Then you heard what I was saying, how bad it is, and you said, I can't win that debate. But you know that if you did win, you would approve that, and that will be almost as bad as NAFTA. Nothing will ever top NAFTA.
Clinton shows again that she conceives the discussion in non-factual terms, and Trump interrupts her once more to indicate that, for him, the disagreement is still factual. Take the following exchange: Thus, while Clinton tries to focus on the general parameters that should determine the economic policy of a government, Trump insists, instead, on two different factual matters -"manufacture is down 30, 40, sometimes 50 percent", and Clinton's political record of support for certain treaties, assumed to be detrimental for the US economy. Both provide clear signs of conceiving the disagreement under different guises, mostly a Type B disagreement for Clinton, mostly a type A one for Trump. It's not our contention that Trump's behavior is inappropriate in this context, while Clinton's is not. Rather, what we contend is that the specific way in which a particular question is turned into a crossed disagreement is facilitated by Trump's reputation to be willing to say anything at any point, to move from one small point to the next, quite unconcerned by the course of the debate or the specific contribution of his interlocutor. Re-framing a debate is not necessarily detrimental for public discourse, at times, the only sensible strategy is to refuse to engage your interlocutor in their own terms. It becomes a problem when public settings that are supposed to contribute to political deliberation systematically turn into crossed disagreements, where both parties end up repeating to their respective supporters the kind of arguments that back their own position. This, we argued above, fosters polarization, by enlarging the size and density of the argument pool for a position.

Case 2
Boris Johnson was interviewed in July 2019 by Sophy Ridge (Ridge, 2019). During the interview, Ridge and Johnson disagreed on whether Johnson has a sense of responsibility towards what he says and does. Ridge inquired whether he felt accountable for the consequences of certain expressions that Johnson had used in the past, in some of his newspaper columns. She provides several examples to show that there is a disparity between what Johnson says and what he does. Again, we underline expressions that point towards a factual conception of the disagreement, and italicized those that we take to be signs of a normative disagreement: RIDGE: You see it's not the first time that something that you've said has got you into trouble is it? I just want to talk a little bit about language that you've used. You've, in columns before that you've written, referred to black people with "watermelon smiles"… JOHNSON: In a wholly satirical way by the way. RIDGE: Can I finish please? "Tank top bum boys", women in burkas who "look like letterboxes". I want to put it to you that you're not homophobic, you've supported gay rights, you've supported a woman's right to wear the burka, but you'll just say anything to get a laugh.
Johnson replies by saying that the criterion for knowing if he is politically responsible is to look at what "he actually did" and anything else would be to take his words out of context. The claim is that under this standard on what should count as "taking responsibility", he: JOHNSON: No, I think if you look at each and every one of those columns or articles you'll find that the quotations have been wrenched out of context, in many cases made to mean the opposite of what was intended, and actually look at my record whilst I was Mayor of London. If you want to look at somebody who's campaigned for gender equality, look at what we did in the Foreign Office where we had a huge campaign for 12 years of quality education for every girl in the world that was massively successful, it was taken up by other countries. Look at the way I ran London. We had huge campaigns to protect the rights of women and particularly against violence against women, and I think sometimes there is a tendency, as I say, to take words out of context rather than to look at what I actually did. Ridge quickly rejects the notion, used by Johnson, of "being responsible" to claim that he is not really taking responsibility for his actions. Johnson, in turn, denies the relevance of the facts Ridge is referring to by offering another set of facts as more relevant to settle the question: RIDGE: It sounds a little bit like someone who just won't take responsibility. You won't take responsibility for the words on Nazanin. You won't take responsibility for what you write in your columns, it was just "wrenched out of context". I mean is that a quality we would want in a Prime Minister? JOHNSON: On the contrary, I take full responsibility for everything I've said. Just look at what I've actually done. When I took over running London eleven years ago, the city was in the grip of the biggest recession that we've known for 50 years. We had riots, heaven knows what we had to go through to get ready for the Olympics. I took full responsibility for all those things. I took responsibility for one thing that was the crime wave afflicting London and I took personal responsibility for it and actually, by dint of some very tough measures, by getting the police out on the street, by giving them the power, the political top cover to do stop and search, we got the murder rate down by 50%, under 100 a year, which is an amazing thing for a city the size of London. Yes, I did take personal responsibility for those things and I take personal responsibility for everything I say and do.
After some symptomatic interruptions from Johnson, Ridge appealed to facts, particularly to some of Boris Johnson's controversial statements that have had negative consequences. Boris Johnson denied that these claims had negative consequences, and also denied that those were the relevant facts to determine whether he had a sense of responsibility. Johnson' dialectical strategy is to question that the facts pointed out by Ridge are relevant to determine whether he takes political and personal responsibility for his statements. In this sense, they are conceiving the disagreement in significantly different terms. Ridge is pointing straightforwardly to some facts that she can use as reasons for her position, while Johnson is trying to reframe the debate by questioning what should be the actual reasons that might be adduced to argue the point. Thus, Johnson appears to conceive the debate as a type-B disagreement, in normative terms; meanwhile Ridge shows signs of conceiving the disagreement as a type-A disagreement, in factual terms, at least at the beginning of the exchange. 6 It is important to notice that, in both cases, they are not talking at cross purposes. They are not talking about different topics, they are rather talking at different levels, as it were. Sometimes one of the subjects in the controversy is talking at a factual level, whereas the other subject is talking at a normative or evaluative level, generating a case of crossed disagreement. This deliberation dynamic is not unusual, in fact it is a common phenomenon of our everyday, deliberative practices. We usually discuss a given topic with other people while moving through different levels of explanation within the debate, often without realizing these conceptual changes. Crossed disagreements are genuine disagreements, and they should be treated as such. In the first example above, Clinton and Trump are arguing about the U.S. economy, and they continue to do so all along the exchange. In the middle of it, Trump acts irreverently, generating (intentionally or not) a crossed disagreement. The normative terms used by Clinton contrast with the factual terms Trump is beginning to use to attack her position. Clinton tries to reframe the debate in order to relocate it in normative terms, but Trump's attitude prevents it.
Crossed disagreements are conceptually and empirically linked to the rise of polarization. As the output of a public debate we expect growing coordination from the subjects in dispute, even if it's just out of the sheer contrast between the relative strengths of the arguments that support each position, but in a real debate like these, as we have seen, the irreverent attitude of one of the interlocutors generates crossed disagreements, promoting an increase in the size and density of the pool of arguments. This exposition to a limited and biased set of arguments tends to make people impervious to the other's reasons (Sunstein, 2017;Barberá et al., 2015;Vicario et al., 2016). They might think that they are taking part in a legitimate and fruitful discussion when, in fact, they are contributing to increase the distance between them.

Conclusion
It has been the goal of this paper to show that a phenomenon that might be reasonably construed by some as a case of testimonial epistemic injustice can be used by certain agents to their advantage to promote their specific political agendas. When public speakers manage to earn a reputation of being somewhat unpredictable or wild, unconcerned 6 Note that in these cases the contending parts do not understand the debate exclusively in normative vs. factual terms. As already pointed above, reallife examples of disagreement are of course heterogeneous: there are no real-life cases of disagreement where a monolithically conceived position (factual, normative, metalinguistic) is confronted with another monolithically conceived position. by the most basic rules of public conversation, this might reduce their credibility overall, they become less reliable, or trustworthy, but it can also be used to their advantage in certain specific contexts. At times, the positive outcome of this policy, presenting themselves as willing to say anything at any time, might outweigh the loss of credibility. The price to pay when not even your supporters expect a well-articulated truthful message from you is potentially high, but the benefits of inhabiting a fringe political persona in many political contexts might also be very high. In particular, in polarized contexts where the mainstream media coverage is perceived to lean towards your adversaries, presenting yourself as someone from whom anything can be expected might provide the room that is needed to turn a strictly directed political debate into a crossed disagreement, a public dispute where the parties show signs of conceiving the debate in different terms. If there are electoral benefits to be obtained by promoting polarization, then galvanizing your supporters might be all you need out of a potentially devastating public appearance, in front of a standard political adversary.
Here's the outline of the argument that we have presented. Under a certain interpretation of what it means to be a member of a disenfranchised group, some agents with a privileged background can reasonably put together a case to argue that they are victims of testimonial injustice. This, nevertheless, is not brought to the public's attention with the intention of correcting this lack of credibility, at least not primarily. On the contrary, having a reputation of being epistemically insouciant is systematically used, consciously or unconsciously, to subvert the goal of certain public debates. Public deliberation is turned into a tool for polarization when a galvanized electorate serves the purpose of those alleged victims of epistemic injustice. Or when they think that it does. By transforming tightly monitored public debates into crossed disagreements, they are able to take advantage of the lack of credibility associated with the public persona that they inhabit to foster polarization. In doing so, they manage to contribute to their political agenda in a situation that was potentially ideal to expose their weaknesses.
Only because the political actors that we are describing can paint their situation as one in which they are victims of an injustice, together with their followers, can they perform as they do in highly controlled public debates. They are given a free-pass by their followers precisely because they are taken to be victims of an injustice. Cultivating a fringe persona does not have the political benefits that they actually obtain out of it unless they are able to portray themselves as victims, as part of a group that already feels victimized. Recognition and identity play a crucial role here, and this is why testimonial injustice is "weaponized". This is what sets apart Boris Johnson and Kanye West. West is not taken to be a victim of testimonial injustice, and he is only taken to be an extravagant person with a mental health problem. Like a racket that is used to hit somebody in the head, a concept being "weaponized" shows nothing with respect to the adequacy of the concept itself, only that we should be careful when we deal with behavior that seemingly enacts this kind of strategy. We can dismiss it, and make fun of it, but we do so at our own peril.