Engaging Epistemically with the Other: Toward a More Dialogical and Plural Understanding of the Remedy for Testimonial Injustice

Abstract

The concept of testimonial injustice (TI) has been expanded considerably since Fricker's groundbreaking original formulation. Testimonial void (TV), as well as other kinds of TI identified in the last decade, encourage the idea that the virtue of testimonial justice (TJ) is not the appropriate remedy to battle against injustice in our testimonial exchanges. This paper contributes to the existing literature on the limitations of TJ as the remedy for TI by drawing attention to its shortcomings in the context of other kinds of TI. By contrast, I propose further engaging epistemically with the other (EE) as a corrective to injustice in our testimonial exchanges. I understand EE as a practice in which information regarding epistemic injustice, strategies to fight it and skills training play the leading roles. If the problem lies in lack of appropriate epistemic interaction between knowers, we need to train ourselves to do what we fail to do. Given that we are in the domain of testimony, EE essentially amounts to engaging further with the other in conversation. In the process, EE sheds light on the need for second-order change to fight TI, as enacting EE would require the transformation of our concepts of testimony and credibility.

1. Introduction: From Self-Correction and Self-alertness to Engaging with the Other

Testimonial injustice (TI) is generally understood as the wrong done toward someone in their capacity as a giver of knowledge. The concept of TI has been expanded considerably since Fricker's (2007) groundbreaking original formulation (cf. Dotson 2011; Medina 2013; Steers-McCrum 2020; Carmona 2021a). This paper examines the concept of the ‘virtue of testimonial justice’ (TJ) under the light of recent research on newly identified kinds of TI. TJ is Fricker's (2007) remedy for TI: a corrective ethico-epistemological virtue specific to the domain of testimony which essentially consists in revising credibility upwards. My point of departure is that if TJ is supposed to neutralize prejudice in our testimonial exchanges, it should be able to do so for all kinds of TI. Reservations about Fricker's guiding ideal of TJ – regarding the difficulties of identifying the unjust credibility judgment as well as of reflating it by the right amount (Alfano 2015; Sherman 2016a, 2016b; Washington 2016) – have already been expressed concerning her original formulation of TI. Building on this work, this paper contributes to the literature on the limitations of TJ as the remedy for TI by drawing attention to its shortcomings in the context of other kinds of TI. In the process, I propose further engaging epistemically with the other (EE) in an active manner as a corrective to injustice in our testimonial exchanges.

In a nutshell, I understand EE as a dialogical practice in which we need to be trained with a view to ameliorating injustice in our testimonial exchanges. By EE, I mean interacting with one's epistemic counterpart in a testimonial exchange in a way that is epistemically relevant for that instance of communication. This might be obtained by further engaging one's counterpart in conversation. One might do it with the purpose of checking whether one's misgivings are justified. However, when one gets into the habit of EE, one's participation might be spontaneous and become second nature.

This paper is primarily motivated by the insights provided by testimonial void (TV) into TJ. In my paper ‘Silencing by Not Telling’ (Carmona 2021a), I present TV as a kind of TI committed by a would-be speaker on a would-be hearer by withholding epistemic materials on the basis on an epistemically and ethically faulty assumption that the would-be hearer lacks the capacity to do anything epistemically relevant with them. The phenomenon of TV is behind inner thoughts such as ‘Why bother telling her? She won't know what to do with it’ (Carmona 2021a: 577). Women's education in Franco Spain is a good example of TV. The intellectual abilities of women were structurally disesteemed during the Franco regime. There was a specific curriculum for girls that prepared them to be (house)wives and mothers, in such a way that “the Franco regime controlled who was (not) told what in the educational context” (Carmona 2021a: 578). Another telling example that I have explored is how women in The Godfather are not told about the family business, on the assumption that they won't understand a man's world. Likewise, I have revisited Fricker's (2007: 90) reading of Anthony Minghella's screenplay of The Talented Mr. Ripley, in which, in the wake of World War II, Greenleaf, owing to gender stereotypes, does not give any epistemic credit to his daughter-in-law-to-be, Marge, who suspects that her fiancée, Dickie, was murdered by a friend of his, Tom Ripley. I have identified as an instance of TV the fact that the male protagonists agree not to tell Marge the sordid facts about her fiancé's life because they assume that she needs protection from the ‘truths of men’.

TV reveals that working on ourselves as listeners won't remedy all instances of TI. Fricker's (2007) original formulation of TJ would have no immediate corrective effect on instances of TV, as the speaker is the wrongdoer and what is undervalued is not someone's credibility but rather their epistemic aptness (Carmona 2021a: 582). Consequently, there is no credibility assessment that needs to be compensated upwards. Rather, remedying TV is a question of improving ourselves as speakers, as givers of epistemic materials. In other words, we need to ameliorate ourselves as sharers of knowledge (Carmona 2021a: 580–2).

Let us return to The Godfather. There is a remarkable shift in the relationship between Michael and his wife Kay once he becomes the head of the Corleone family (Carmona 2021a: 578). Initially, Michael has a distant attitude toward his own family, as he clearly disapproves of the family business. At that point, he trusts Kay and does tell her about the unorthodox methods followed by his father to support his godson Johnny Fontana's singing career. Kay's reaction is indicative of her competence to deal with the information. Despite being repelled by the family's way of doing business, she is sympathetically aware of her husband's mixed feelings about his family. Nevertheless, after taking over his father's business, Michael starts to conceal information from Kay, who is systematically treated from that moment onwards as someone who cannot handle ‘a man's world’. For instance, Michael withholds from Kay that he is again involved in the most morally distasteful affairs after years trying to ‘legitimize’ the family business – when a virtuous speaker would have told Kay, especially as their daughter played a role in the legitimate part of the business.

Imagine a different end for the trilogy aimed at compensating the secondary role that women play in the story. Suppose that, on the night when their son Anthony appears in Cavalleria Rusticana in the Teatro Massimo in Palermo, Michael does tell Kay about the drastic turn taken by the family business. Consequently, after the show, she makes their children leave the theater with her through a backdoor. As in the film, Michael is shot on the Teatro Massimo stairs and survives. However, in our hypothetical ending, it is his nephew Vincent who receives one of the bullets directed at Michael and dies. Kay's sensible precaution prevents their daughter's death. Michael realizes that if Kay had not acted as she did upon receiving the information, one of their own children could be dead. Her action is thus indicative of her insights into the workings of ‘a man's world’. This is an example of how, if Michael had engaged with Kay in conversation, Kay's epistemic aptness could have helped Michael fight the prejudice that women had to be kept away from certain epistemic materials.

Likewise, if Greenleaf would have engaged with Marge in conversation, instead of concealing information from her, he could have realized that Marge didn't need protection from the unpleasant facts concerning his son's life. In fact, Greenleaf could have realized that Marge was already aware of that dimension of her fiancé's life, as on her view their bond lied precisely in her ‘understanding’ attitude. Accordingly, by engaging with Marge in conversation, Greenleaf could have come to the realization that he did not need to withhold information from her. In other words, engaging epistemically with Marge would have prevented Greenleaf from committing TV on her. This is another example of how EE creates friction to neutralize the effect of prejudice on one's judgment.

By the same token, if individual teachers had engaged epistemically with girls in Franco Spain before withdrawing from them certain epistemic materials, the competence that girls would have manifested could have contributed to weakening existing prejudice regarding the intellectual abilities of women, in such a way that structural TV would have been debilitated. By contrast, the fact that women were generally not exposed to certain epistemic materials (i.e. the deliberately constructed gaps in their education) reinforced existing prejudices regarding their epistemic aptness, as the educational system undermined their full development as epistemic agents (Carmona 2021a: 584; Carmona 2021c: 10–11).

EE can also remedy instances of Fricker's original formulation of TI. Consider that a physics teacher, Magan, is evaluating a presentation of a group of students. The activity is oriented toward fostering teamwork. Magan specifies that students ought to engage in discussion and produce a unitary piece of work, in such a way that each student has a comprehensive understanding of the whole. After listening to the presentation, Magan has the impression that Neetu, one of the students in the team, lacks an overview of the work as a whole. Past experience does not work in Neetu's favor. Magan has the overall feeling that Neetu often seems distracted and uninterested. However, the truth is that Neetu is the only female student in physics class and that Magan has the prejudice that girls are not really interested in physics. We would be before a case of TI, as Magan would have given Neetu's presentation less than merited credibility owing to the influence of prejudice.

Now suppose that, instead of giving Neetu a lower grade than to the rest of her team straight away, Magan decides to engage Neetu in conversation with the purpose of seeking further evidence regarding her comprehension of the task. Magan might suspect that some form of prejudice is influencing his judgment, as he cannot identify explicitly what is that Neetu did wrong. However, prejudice awareness does not need to be part of the picture. Magan might simply be a rigorous grader and like to be able to explain his motives to give a specific grade in case students ask. For that purpose, Magan asks Neetu a question that to be answered requires the perspicuous understanding aimed at by the task. Neetu provides a satisfactory answer, and Magan realizes that he was misjudging her. Accordingly, Magan decides to engage with each student presenting in a team to check whether their individual contribution and understanding is solid enough. In other words, Magan develops the strategy of engaging with each single student before grading their contribution to teamwork.

My proposal is also inspired by Fricker's insight that we must work on our attitude toward the other to be in an appropriate position to judge. We learn from Fricker (2007: 79–80) that our credibility judgments require sufficient empathy. For building empathy, engagement is fundamental, and the more interaction the better. By interacting with one's epistemic counterpart, an epistemic agent might ultimately develop an attitude toward the epistemic other that comes into conflict with prejudiced dispositions to act. Fricker understands exposure to the epistemic other as an effective medicine against prejudice:

Yet with the degrees of familiarity – gained over the duration of a conversation, or perhaps a more sustained acquaintance – the prejudiced first impression melts away, and the hearer's credibility judgement corrects itself spontaneously. (Fricker 2007: 96)

As discussed in sections 3 and 5.2 especially, the kind of epistemic engagement with the other that I propose aims at obtaining the degree of familiarity that can be obtained over the duration of a conversation with a view to strengthening the kind of epistemic trust we owe one another as human beings. For instance, the gesture of greeting a passer-by before asking them for a piece of information might have a positive effect on the epistemic exchange. Similarly, asking if one can be of any more help after sharing a piece of information might have a beneficial impact on the uptake that such epistemic materials receive.

Besides obtaining familiarity and strengthening epistemic trust, when EE, we also obtain further evidence on which to base our epistemic assessments. Recall that by asking Neetu the appropriate question, Magan realizes that he has misjudged Neetu's epistemic contribution. Besides fighting prejudice, EE can also be helpful to neutralize evaluations owing to epistemic vices other than prejudice. Suppose that Magan was an unprejudiced yet careless teacher, because of which he has underestimated Neetu's epistemic contribution. Active epistemic engagement with Neetu could make Magan realize his mistake as well as the need to pay more attention when dealing with students.

While discussing TJ, Fricker leaves the virtuous hearer on her own. Her virtuous hearer is someone whose testimonial sensibility has been reconditioned by past anti-prejudicial self-corrections and who retains ongoing self-alertness toward the impact of unfamiliar prejudice on their credibility judgments. One would expect that other epistemic agents had some kind of impact on both the transformation and the watchfulness. However, in Fricker's model, other epistemic agents have no relevant role in reconditioning one's testimonial sensibility (even though it would be consistent with her view). Against this background, I subscribe to a collaborativist solution in claiming that a drastic shift away from individualism is necessary in the context of TI (Washington 2016).

A fundamental feature of my proposal is the vindication of the role of the one at the receiving end of the epistemic injustice in reconditioning the perpetrator-to-be's testimonial sensibility. In the process, I depict an epistemic subject who may be acting virtuously, but is not committed to any strong ideal of motivational self-sufficiency of character. EE also sheds light on the need to stop conceiving our testimonial exchanges as single transactions in which a hearer is supposed to receive epistemic materials from a speaker. By contrast, testimony ought to be understood as a product of the active epistemic engagement between epistemic agents who might interchange roles, in such a way that the speaker becomes hearer and vice versa. Consider again the interaction between Magan and Neetu. After listening to Neetu's presentation, Magan becomes the speaker by asking Neetu a question before returning to his role as hearer when Neetu answers. Accordingly, both Magan and Neetu have flexible, interchangeable roles in the testimonial exchange.

In what follows, section 2 outlines Fricker's original proposal. Section 3 shows that TV forces us to reconsider the viability of TJ as the most effective remedy for TI broadly understood and explores the two essential purposes of EE, namely gathering (and delivering) evidence of epistemic competence and strengthening epistemic trust between knowers. Section 4 shows how the angle to TJ explored in the previous section is also ameliorative concerning kinds of TI other than TV, such as ‘self-appointed speaking for’ (Steers-McCrum 2020), Medina's (2011) TI owing to credibility excesses, and Dotson's (2012) testimonial quieting and testimonial smothering. The reader will find an overview of EE in section 5. EE is portrayed as a practice with the two essential purposes mentioned above in which information regarding epistemic injustice, strategies to fight it and skills training play the leading roles. This section also reveals that fighting TI requires first- and second-order change and that EE leaves plenty of room for the virtuous agent. Finally, section 6 explores some of the benefits and limitations of EE without intending to be exhaustive. EE is said to foster epistemic democracy and to leave less room than TJ for Fricker's (2007: 103) concept of moral and epistemic bad luck. It also explores the negative impact that oppressive power relationships have on EE. Though for reasons of space I cannot deal as much as it needs to be done with the issue of how proper EE is able to be enacted within the frameworks of structural epistemic injustice, I do consider it in section 6.3, while acknowledging the limitations of my discussion.

2. Fricker's Original Formulation of TI and TJ

Following Fricker (2007), testimonial injustice (TI) is generally understood as the wrong done toward someone in their capacity as a giver of knowledge. TI occurs when prejudice on the part of the hearer leads to the speaker's testimony receiving less than merited credibility. Fricker's central case of TI is due to identity-prejudicial credibility deficit. In its more severe forms, TI is both persistent and systematic. In other words, it has a diachronic dimension, as it develops and evolves through time, and a synchronic one, which accounts for how it is harmoniously intertwined with other forms of injustice by a common prejudice at a given time (Fricker 2007: 29). The main fuel for TI is identity prejudice, that is, stereotypes regarding social identities that are constitutive parts of the social imagination. Epistemic agents can be under the influence of identity prejudice even if it is inconsistent with their beliefs (Fricker 2007: 37).

Fricker (2007: 27–8) acknowledges that identity prejudice can manifest itself in positive or negative form, that is, as prejudice for or against people owing to their social identity. However, she is mainly concerned with its negative form because of her focus on credibility deficit. This focus is explained by her understanding of TI as token cases. By contrast, she argues that credibility excess only harms the subject if it is cumulative (Fricker 2007: 21). Consequently, she uses “identity prejudice” as an abbreviation for “negative identity prejudice”.

Fricker suggests compensating upwards for the influence of identity prejudice by means of the virtue of testimonial justice (TJ): “a virtue such that the influence of identity prejudice on the hearer's credibility judgment is detected and corrected for” (Fricker 2007: 5–6, my emphasis). In essence, TJ is “a matter of one's credibility judgements being unprejudiced” (Fricker 2007: 92–3). As TJ is difficult to obtain, sometimes, given the nature of prejudice, the right thing to do is to suspend judgment. In unclear cases in which one is forced to decide, she suggests seeking further evidence (Fricker 2007: 91–2). TJ can be displayed naïvely, in such a way that one is prejudice-free from the start, or in corrective form. In corrective form, TJ can be exercised reflectively or in a more spontaneous way. Familiar prejudices might be neutralized by habituation, in such a way that the reflective requirements involved become second nature. However, dealing with less familiar prejudices is likely to require ongoing active critical reflection (Fricker 2007: 98).

This brief presentation of TJ should suffice for the time being. In the course of addressing other forms of TI in sections 3 and 4, TI and TJ will be further characterized. In the process, we shall see that EE can be a good remedy to fight such kinds of TI.

3. Testimonial Void

We saw in section 1 that TV is a newly identified form of TI committed by a would-be speaker on a would-be hearer by withholding epistemic materials on the basis of an epistemically and ethically faulty assumption that the would-be hearer lacks the capacity to do anything epistemically relevant with them. In what follows, we shall see that the extension of the limits of the concept of TI entailed by TV brings more insight into the remedy we need to correct for the influence of prejudice in our testimonial exchanges. We shall elucidate whether TJ is effective (i) if what is assessed is epistemic aptness instead of testimonial credibility, and (ii) if the injustice is committed by the would-be speaker.

Let us look more closely into Fricker's depiction of TJ in the context of the virtuous hearer. When suspecting prejudice in her credibility judgments,

The guiding ideal is to neutralize any negative impact on prejudice in one's credibility judgments by compensating upwards to reach the degree of credibility that would have been given were it not for the prejudice. And this is at least an ideal that can regulate our practice of credibility judgment. The upshot of approximating this ideal in any given case may indeed be a reflation of the credibility level in the judgment; or, alternatively, it might be that all that we are able to do is to render our judgment more vague and more tentative. In cases where the whole business of judging credibility becomes too indeterminate, we may need to suspend judgment altogether; or, alternatively, if we have a responsibility to arrive at some definite verdict, we may have resources to seek out further evidence. At any rate, since it is not always possible to ‘correct for’ the impact of prejudice by making a neat compensatory reflation of credibility, let us say that, one way or another, the virtuous hearer neutralizes the impact of prejudice in her credibility judgements. (Fricker 2007: 91–2)

Accordingly, Fricker makes explicit that the possession of “the anti-prejudicial virtue that she is looking for”, namely TJ, “requires the hearer to reliably neutralize prejudice in her judgements of credibility” (Fricker 2007: 92). As we saw in section 1, neutralizing prejudice in our credibility assessments would have no impact on instances of TV, as it is not credibility that is assessed in a biased manner. As epistemic aptness includes epistemic credibility, one could reformulate TJ as the anti-prejudicial virtue that neutralizes prejudice in one's judgments of epistemic aptness. However, this is so vague that it is not clear how we should proceed. Besides, this reformulation of the virtue would suffer at least from the same difficulties as TJ in respect of identifying the unjust credibility judgment and reflating it by the right amount (Alfano 2015; Sherman 2016a, 2016b; Washington 2016).

For instance, we learn from Sherman that it is likely that we remain unaware of our blind spots, especially in the case of implicit bias. Most of your views seem right to you, in such a way that

reflecting on testimonial justice might bring to mind a few errors, but you are likely to think the vast majority of the time, your judgements are fair and accurate, otherwise, they wouldn't persist in being your judgements. (Sherman 2016a: 238)

By definition, implicit bias is inaccessible to the biased epistemic agent's consciousness (Pronin et al. 2002). We might fail, therefore, at the detection stage. As a result, we are in a difficult position to recognize when to implement the policy of reflating attributed credence, or, for that matter, by how much to reflate. In fact, under-correction is the norm (cf. Epley and Gilovich 2006).

Moreover, TV is by itself particularly difficult to detect, as there is no actual epistemic exchange (Brick 2021: 48; Carmona 2021b: 1–5). In fact, the victim of the injustice is a would-be hearer, that is, someone who has not participated in the testimonial exchange. By contrast, in instances of Fricker's original formulation of TI, the piece of testimony coming from the marginalized speaker might function as evidence of their epistemic competence and consequently create friction with prejudiced dispositions to act (Carmona 2021a: 586).

In addition, TV draws our attention to the fact that action needs to take place not only at the level of assessment, but also at the level of engagement with the other, as its basic form entails that a would-be-speaker withholds epistemic materials. It might sound uncharitable to say that Fricker (2007) remains in the field of judgment. In a sense, it certainly is, as the purpose of her work is to have an impact on our understanding of our epistemic practices, so that they are transformed for the better. In fact, one could say that judging is already doing something, acting in certain ways, for instance, drawing from Wittgenstein (1975 [1969]: §§ 232, 395). That said, my feeling is that Fricker's formulation of TJ does not draw our attention to action as much as it is necessary for the issue at hand. I think that this is related to the fact that she chooses virtue-theory as the solution to epistemic injustice, in particular a rather individual-oriented Aristotelean conception of virtue (cf. Davidson and Kelly 2015; Sherman 2016a, 2016b). However, discussing this point further goes beyond the scope of this paper.

Fricker acknowledges that reflation of one's assessment is not always possible. Accordingly, she identifies other forms that TJ can take. Three alternative options are mentioned. We are to prefer option one:

1. (i) rendering our judgment more vague and more tentative.

When judging credibility becomes too indeterminate, she recommends

1. (ii) suspending judgment altogether,

or, alternatively, when there is ineludible responsibility to “arrive at some definite verdict”,

1. (iii) seeking out further evidence.

Rendering our judgment more vague and more tentative might not have the right effect on instances of TV either. Vagueness and tentativeness are unlikely to push us to share epistemic materials. In fact, when there is no manifest item to be judged, such as a piece of testimony, indetermination might be overwhelming. As explained, TV always takes place in what could be portrayed as a pre-emptive form, as the epistemic agency of the one at the receiving end of the injustice is rejected before manifesting itself (Carmona 2021a: 585). Accordingly, the very nature of TV, as the would-be-hearer's epistemic qualities might remain latent, makes TJ harder to obtain.

Fricker's first option in case of overwhelming indetermination is no satisfactory remedy for TV either, as reserving judgment has the same effect as the epistemic injustice in question: the would-be speaker does not contribute to the epistemic exchange, withholding the epistemic materials they have to offer. Reserving judgment does not alter the status quo. Rather, it reinforces it, as speaker and hearer remain as isolated from each other as ever.

In fact, my feeling is that reserving judgment is no satisfactory solution either in case of TI. Reserving judgment might entail displaying the same attitude toward the epistemic materials that one has to assess than simply undervaluing them, for what takes place when one does not give merited credibility is that one does not proceed as if such materials were true. For instance, Greenleaf did not take actions against Ripley upon hearing Marge's suspicions. Likewise, when one reserves judgment upon receiving certain epistemic materials, one does not act as if they were true.

I have never understood why seeking further evidence is Fricker's third option. As I see it, it is her best option. I believe that the dissimilarities in our approaches might be a consequence of understanding differently the form that seeking further evidence should take. If one has in mind that seeking further evidence entails enquiring about and navigating one's counterpart's epistemic past record, I agree that it is neither the easiest nor the best way to proceed. Though it is something that could be done to build a case in a court of law, it is rare that we have the option to do that in everyday testimonial exchanges.

By contrast, I suggest that the epistemic assessor seeks further evidence by means of engaging epistemically with the other epistemic agent. EE might avoid the act of withholding information altogether, as one is supposed to approach the would-be hearer. For instance, one could double-check judgment by sharing epistemic materials related to the information that one is doubtful that the other epistemic agent is apt to receive. However, if we get into the routine of EE, double-checking might often be unnecessary. Understanding testimonial exchanges as activities in which everyone involved have flexible roles to play might directly lead to more just interactions.

EE can take different forms. For example, by asking relevant questions to the would-be hearer, one might be able to make sure whether the would-be hearer is epistemically ready for the epistemic materials one has to offer. Recall our earlier examination of the case of Marge in The Talented Mr. Ripley. Given that Marge was indeed ready to learn about the sordid facts regarding her fiancé's life, if Greenleaf had sounded her out instead of biasedly assuming that women are not ready for ‘the truths of men’, he could have ended up telling her.¹ In this regard, to make sure that Marge was ready to receive such epistemic materials, Greenleaf could have approached the issue at hand indirectly, as if he was talking about someone else. Likewise, delivering gradually to Marge the tawdry information concerning her fiancé could be a good strategy to check whether she could bear the epistemic materials coming next. Both strategies could also be combined.

The gradual delivery of information is often employed in educational contexts. Especially when teaching a difficult topic, teachers deliver information by degrees, checking at different stages whether students are ready to move forward. This is also true in the case of feedback. If a student performs badly in an exam, the more sensible thing to do is to handle one problem at a time, so that they don't feel overwhelmed. If the speaker feels at some point that the would-be hearer is saturated with information, they might non-culpably postpone giving a certain part of their feedback. As a matter of fact, this is a virtue in a teacher.

Engaging with the would-be hearer might also contribute to the defeat of prejudice by fostering empathy. A would-be speaker, like the hearer in the classic conception of TI, needs to empathize sufficiently with the would-be-hearer with a view to fighting against identity prejudice. Fricker explicitly argues that our credibility judgments require sufficient empathy, so that we are in an appropriate position to judge a speaker, “and empathy typically carries some emotional charge” (Fricker 2007: 79). She portrays emotion as capable of cognitive content. Fricker advances a model for non-inferential judgment of testimony that consists of five principles. Her conception of emotion articulates principles 3 (judgment is intrinsically motivating), 4 (intrinsically reason-giving) and 5 (contains an emotional cognitive aspect). Fricker's insight is that when a hearer feels that someone is telling the truth, they are motivated and justified to believe the speaker's word. As a result, “the feeling of trust in the virtuous hearer is a sophisticated emotional radar for detecting trustworthiness in speakers” (Fricker 2007: 80).

Likewise, sufficient empathy is needed to be in an appropriate position to judge a would-be-hearer's epistemic aptness when identity prejudice is at work. In the absence of counterevidence, empathy might contribute to the generation of a feeling of trust that might lead to fairer assessments of epistemic aptness. And for building empathy, engagement is fundamental. In fact, by interacting and engaging with the hearer, the speaker might ultimately develop an attitude toward the hearer that comes into conflict with earlier dispositions to act. Building this attitude entails the more interaction the better. For example, the classroom is a right environment in which to nourish such an attitude. Empathy is less likely to emerge in briefer epistemic interactions in less controlled environments. However, the experience of developing such a feeling of trust toward our epistemic counterparts in certain cases might have a positive impact on how we generally approach others in our everyday epistemic exchanges, even in the most harried and abrupt cases. In fact, if we were trained in EE from early childhood, participation and enaction of the attitudes that it entails could become second nature.

Recall that TV might be unprejudiced. Other epistemic vices, such as arrogance or carelessness, can lead to TV. Accordingly, TJ wouldn't correct for such unprejudiced instances. The next section explores TJ's limitations when trying to ameliorate unprejudiced instances of TI.

4. TJ and EE under the Scrutiny of Other Kinds of TI

4.1. Self-appointed speaking-for

The phenomenon of self-appointed speaking-for (SASF) “occurs when one speaks on behalf of or in place of another individual or group without their authorization” (Steers-McCrum 2020: 241). We are given as a paradigmatic case the following example:

You are on a date at a restaurant you have never been to before and are still considering the menu when the server comes to take your order. Your date gives his order and the server turns to you, “And for the lady?”

Before you can say anything, your date answers for you: “She'll have …” (Steers-McCrum 2020: 241)

SASF extends the category of TI for at least two reasons. On the one hand, the speaker commits the injustice by speaking for their counterpart. On the other, SASF can be due to epistemic vices other than prejudice, such as arrogance or carelessness. In fact, no epistemic vice needs to play a role in SASF, as “the act itself is disrespectful and exclusionary” (Steers-McCrum 2020: 242).

We learn from Steers-McCrum (2020: 244) that one could understand SASF as a variety of pre-emptive testimonial injustice (PTI). Structural prejudice in PTI “does its work in advance of a potential testimonial exchange”, preventing “any such exchange” (Fricker 2007: 130). Likewise, SASF “silences and blocks victims from exercising their capacity to make epistemic contributions and their ability to control which and how contributions are made in their names” (Steers-McCrum 2020: 244).

Given that the possession of TJ basically amounts to the hearer reliably neutralizing prejudice in their credibility assessments (Fricker 2007: 92), TJ wouldn't be able to remedy those instances of TJ owing to epistemic vices other than prejudice or those in which epistemic vice plays no role. In addition, like in TV, as the speaker commits the injustice, reflating one's credibility assessment by the right amount won't remedy SASF. Besides, as seen in section 3, the misgivings about the success of TJ as a remedy for TI increase when the victim of the injustice does not have a chance to exercise their epistemic agency (Carmona 2021a: 586, 2021b: 1–5). The fact that certain groups of people are never asked for information does not make neutralizing prejudice any easier for the virtuous hearer. Likewise, the fact that in SASF the victim does not get to speak might make the phenomenon particularly difficult to detect as a form of TI. After all, there is no piece of testimony coming from the would-be-speaker to be interpreted as a trace of the epistemic injustice.

By contrast, by EE, SASF could be avoided in all possible instances of the phenomenon. Observe that EE is precisely what is missing in SASF, as the speaker-for does not let the silenced speaker contribute to the testimonial exchange. Someone trained in EE, at the very least, would ask their date in the restaurant what they want to eat, instead of speaking for them. Familiarity with EE might also lead the epistemic agent in question to remain silent and welcome the would-be speaker's contribution. In the case of prejudice, our interaction with our epistemic counterpart is likely to create friction between their epistemic competence, actualized in each instance of their participation, including those in the role of speaker, and our discriminatory inclinations. Accordingly, our trust toward them as competent speakers is likely to be strengthened.

4.2. Credibility excess

Medina (2013: 60) is unconvinced by Fricker's theses that credibility excesses

1. (i) “can only qualify as a special case of cumulative injustice, but not as a regular case of testimonial injustice”, and

2. (ii) don't “harm the speaker in the course of the exchange in the same way that a credibility deficit”.

By contrast, Medina (2013: 60) claims that “some of what Fricker considers the long-term effects of credibility excess can actually appear in the course of a conversation”. He adds that credibility excesses might entail “an epistemic harm that affects all involved in the testimonial exchange – speaker, hearer-attributor, and other interlocutors included” (Medina 2013: 61). For instance, a speaker subjected to credibility excess might develop epistemic vices, such as arrogance or dogmatism, “if the exchange is complex enough or goes on for long enough” (Medina 2013: 60). Regarding other interlocutors (including the speaker), the speaker's self-assured epistemic voice might intimidate them. Consider that a hearer suffering from an inferiority complex wrongs themself “by attributing a credibility excess to all those who are different from” them “and credibility deficits to all that are like” them (Medina 2013: 60–1). Other interlocutors or witnesses of an excessive assessment of credibility might also be harmed when exposed to such a disproportionate attribution of epistemic trust.

We learn from Medina (2013: 61) that “credibility has an interactive nature”, in such a way that it is “comparative and contrastive”: “being judged credible to some degree is being regarded as more credible than others, less credible than others, and equally credible than others”. This raises a crucial point: “credibility never applies to subjects individually and in isolation from others, but always affects clusters of subjects in particular social networks and environments” (Medina 2013: 61). Accordingly, credibility excesses affect the very dynamics of our testimonial exchanges.

Could TJ help in this context? Would neutralizing (in this case, positive) identity prejudice help? At the very least, one would have to deflate instead of reflating one's credibility assessment to compensate for the effect of prejudice. However, it is as dubious as in the case of TI as understood by Fricker (2007) that one could compensate for the right amount. Acknowledging that compensating for the right amount is tricky while discussing Fricker's original formulation of TI, Davidson and Kelly (2015: 18) suggest over-reflation as a form of restorative justice to those who are marginalized. However, over-correction is at least as counterproductive as TJ (cf. Sherman 2016b). Besides, erring in the context of TV might be more dangerous than in the context of TI. In cases of credibility excess, if one over-corrects, one might end up committing TI owing to credibility deficit.

Furthermore, detecting for prejudice in cases of TI owing to credibility excess might be more difficult than in cases of credibility deficit. As an inferiority complex might often be behind credibility excess, besides fighting prejudice, one would have to fight the negative effects that the inferiority complex has on one's epistemic agency. We also need to consider that in cases of credibility excess, one might be under the influence of positive identity prejudice regarding the epistemic other and negative identity prejudice in respect of oneself. The interaction between the two can lead to a vicious circle:

my credibility excess towards others can have the effect of my voice feeling inhibited, my becoming vulnerable to gullibility, my self-trust shaken or fading in comparison to the disproportionate epistemic trust given to the speaker, and so on. (Medina 2013: 61)

A shaken self-trust is likely to make (both positive and negative) identity prejudice harder to detect. Likewise, if correcting for prejudice is hard enough when we do not suffer from such ills, it certainly becomes a very hard task to accomplish when we do. Besides, if one considers that credibility excess might be a result of epistemic vice other than prejudice, such as carelessness or underconfidence, TJ is unlikely to be a good remedy for this kind of injustice, as prejudice may not exist.

By contrast, EE might be a good way to detect a reflated credibility assessment, as in the interaction with the other one might come to see that such a person does not deserve the authority that we are disproportionally attributing to them. For instance, even carelessness might be easier to handle in a long enough testimonial exchange, as one is exposed to more evidence. Consider the possibility that one's credibility assessment is reflated and that at a later time there is sufficient friction between new evidence and one's initial credibility assessment. Things might be more complex if we are underconfident, as it has a direct effect on one's self-trust. However, EE might make possible that by interacting with the other one realizes that one is treating oneself wrongly if the differences between one's own epistemic aptness and that of one's counterpart are not disproportionate.² In fact, by exercising our own epistemic agency and realizing that we do not fall short of the mark, our self-trust might be strengthened.

EE might also help to correct credibility excess owing to prejudice. As discussed in the context of underconfidence, with more evidence regarding one's counterpart's epistemic aptness, including (yet not restricted to) credibility, friction between evidence and prejudice might act as a corrective and motivate us to reconsider our tendency to overestimate someone's credibility.

4.3. Other kinds of epistemic injustice

Though discussing the benefits of EE for fighting other kinds of TI goes beyond the scope of this paper, let me add a few remarks regarding testimonial quieting (TQ) and testimonial smothering (TS), as they receive a good deal of attention in the literature on epistemic injustice and we shall return to TS in our discussion of the limitations of EE owing to oppressive power relationships.

TQ and TS are two practices of silencing. Dotson (2011: 241) understands a practice of silencing as “a repetitive, reliable occurrence of an audience failing to meet the dependencies of a speaker”. This kind of violence owes to pernicious ignorance, “a reliable ignorance or a counterfactual incompetence that, in a given context, is harmful” (Dotson 2011: 242), as a result of which, epistemic agents fail to track certain truths, causing harm.

In this regard, TQ “occurs when an audience fails to identify a speaker as a knower” (Dotson 2011: 242). Drawing on Collins (2000), Dotson gives the example of black women in the US, who are systematically undervalued as knowers. For its part, TS takes place when a speaker truncates their own testimony “in order to insure that the testimony contains only content for which one's audience demonstrates testimonial competence” (Dotson 2011: 249), in particular regarding unsafe and risky content. To explain what she means by ‘unsafe and risky’ epistemic materials, Dotson (2011: 244–5) refers the reader to Kimberlé Crenshaw's discussion of the silence around gender violence in certain non-white communities, as such information could reinforce harmful stereotypes.

If EE became widespread and habitual, TQ would become less likely to occur, as EE increases opportunities for contact and individuates knowers, in such a way that we would be exposed to the epistemic aptness of marginalized individuals. Therefore, there is more chance that members of marginalized epistemic groups are recognized as knowers. Likewise, with more exposure to the realities of marginalized identities, one is likely to be more aware of the need to manifest willingness and testimonial competence for a speaker to feel safe to share unsafe and risky epistemic materials. Accordingly, the kind of ignorance adding fuel to TS is likely to be undermined. In consequence, (already) less marginalized knowers are likely to feel safer to trust non-marginalized individuals with their pieces of knowledge. By contrast, remedies aimed at changing the individual agent without exposure to marginalized identities, such as TJ, are unlikely to be helpful to ameliorate practices of silencing such as TQ and TS.

5. Engaging Epistemically with the Other: An Overview

5.1. EE as a practice

TV, as well as other kinds of TI identified in the last decade, encourage the reservations regarding TJ as an appropriate remedy to battle against injustice in our testimonial exchanges (Alfano 2015; Sherman 2016a; Washington 2016). I share Sherman's (2016a: 245) views that “the will to be just is probably much less important than the information and strategies involved in identifying epistemic injustices” and that “thinking about testimonial justice as a virtue seems to be no help at all in avoiding epistemic injustice”. By contrast, such recently identified kinds of TI call for reconsideration of the relevance and benefits that the active engagement with the other might have for our everyday epistemic interactions.

We learn from Doris (2015) that certain individualist conceptions of agency are untenable. Most of our views about the world are formed in conversation, therefore, it should be a matter of concern that changing individual reasoners has been the focus of our approaches to tackle epistemic injustice (Washington 2016). In this regard, EE can be understood as a collaborativist solution that responds to the insight that “promoting epistemic justice could be achieved not only by changing the individual agent but also by ameliorating the material or social context” (Alfano 2015: 61). In this regard, EE aims at ameliorating the dynamics of testimony, so that we understand it as a dialogical activity in which all the parties involved have flexible roles to play.

I understand EE as a practice with two essential goals, namely gathering (and providing) further evidence of epistemic competence and fostering a feeling of trust among knowers, in which information regarding epistemic injustice, strategies to fight it, and skills training play the leading roles. If the problem lies in lack of appropriate epistemic interaction between epistemic agents (a hearer who does not give proper uptake to received testimony, a speaker who does not share epistemic materials, a hearer who fails to show testimonial competence, and so on), the practice in which we need to be trained is precisely EE. It is as simple as that. We need to train ourselves to do what we fail to do. Given that we are in the domain of testimony, EE essentially amounts to engaging further with the other in conversation.

With a view to detecting injustice in our testimonial exchanges, education from early childhood ought to include information concerning identified kinds of TI. Awareness of the existence of injustice in our testimonial exchanges and knowledge of specific cases and related-concepts ought to be understood as a pillar of civic learning. For instance, with a view to ameliorating TS, one would need to be familiar with the concept of TS and related concepts, such as the notion of ‘unsafe and risky’ content, as well as with paradigmatic examples, if possible, to make it more tangible, from one's own social context.

In order to achieve its two goals, one could say that EE always puts into practice at least two core strategies. By engaging with one's epistemic counterpart,

1. (i) one increases opportunities for contact and positive epistemic interactions, and

2. (ii) one individuates them, in such a way that one obtains specific information from them as an individual.

With more contact, the opportunities for gathering (and delivering) evidence of epistemic competence increase and this might in turn foster or strengthen the feeling of trust toward one another that we need for our epistemic interactions to function smoothly. Likewise, with more interaction with one's epistemic counterpart, we are likely to be in a better position to individuate them, which is likely to have a positive effect on the quality and the quantity of future interaction. In addition, obtaining information from them as an individual might in turn have a positive effect on our feeling of trust toward that person as a knower. In other words, the two core strategies strengthen each other and are useful for the two essential purposes of EE.

Regarding the success of the core strategies of EE in battling against bias, recent research on strategies for mitigating implicit biases (Devine et al. 2012, 2017; Forscher et al. 2017) show that both individuation and incrementing opportunities for contact reduce the influence of implicit bias. Firstly, individuation, which consists in “preventing stereotypes inferences by obtaining specific information about group members”, leads to evaluations of “members of the target group based on personal rather than group-based attributes” (Devine et al. 2012: 11). Secondly, “increased contact can ameliorate implicit bias through a wide variety of mechanisms, including altering the cognitive representations of the group or directly improving evaluations of the group” (Devine et al. 2012: 11–12).

We also need training in paradigmatic strategies of epistemic engagement to fight specific instances of TI. Which strategy is best suited for a specific instance of TI will always be context-dependent. However, direct acquaintance with basic strategies will facilitate the task of battling specific instances of TI. Some basic strategies have already been mentioned in this paper, especially in sections 3 and 4. For example, concerning the gathering of further evidence in respect of one's counterpart's epistemic competence, we learnt from our discussion of TV that one could check whether they are ready to receive certain epistemic materials by sounding them out; for instance, by gradually delivering information so as to make sure they are able to deal with the epistemic materials coming next. Likewise, in respect of Fricker's original formulation of TI, we could double-check whether a piece of testimony is trustworthy by asking our fellow epistemic agent a relevant question. As for avoiding TS, it is us who should provide further evidence regarding our testimonial competence. For that purpose, we should be trained in showing testimonial competence whenever risky content is involved; for instance, by making explicit when appropriate that one knows about the specific difficulties suffered by a community. Public policy ought to be oriented toward refining such strategies as well as toward designing the appropriate training programs to make sure citizens employ them well. We need skills training. We will single out some of the skills that we need to acquire in the next section.

5.2. Core Skills for EE

This section homes in on the details of what effective EE could look like in practice by drawing from the literature on critical thinking that deals with debates about social identity and social difference in which stereotypes and biases unfortunately play a big role. In particular, we will look closely into Linker's (2015) notion of ‘intellectual empathy’. On her reading, intellectual empathy entails a combination of “knowing about ourselves and knowing as much as we can about other people's circumstances, particularly people whose circumstances are different from our own” (Linker 2015: 13). This means that we need to be informed, but reliable information in this context takes a very specific form: we need to look “at the situations people face through their eyes” (Linker 2015: 13), which “requires critically and creatively imagining how that information is understood and processed by people whose experiences are different from our own”, as well as “how we ourselves, with our own particular social circumstances, are seen and understood by people whose social identities are different from our own” (Linker 2015: 14).³

We find plenty of real-life examples in These Walls Between Us, in which Wendy Sanford (2021), a white woman co-author of the women's health classic Our Bodies, Ourselves (Boston Women's Health Book Collective 1970), reflects on her still-evolving sixty-year connection with Mary Norman, a black woman who, since the age of fifteen, worked for her wealthy family as a live-in domestic worker during summer vacations. Sanford shares with her readers some of the occasions in which she tried to put herself in Mary's shoes as well as many of those in which she failed to do so.

She tells us about how she imagines a 15-year-old Mary travelling for seven hours with a white, off-duty cop hired by Sanford's mother, Nina, to drive Mary safely from Princeton to the ferry that would take her to Nantucket (Sanford 2021: 10–11). The fact that Nina did not consider that Mary might have been scared of being alone with him, given that white police officers enforced the Jim Crow where Mary came from, is not the only example of white solipsism that Sanford identifies in her mother's attitudes toward Mary; for instance, we are told that Nina assumes that Mary could swim and that she would feel free to do so in a beach full of wealthy white people (Sanford 2021: 18–19). As for her own failures, Sanford reports that she failed to listen to Mary and put herself in her place when Mary began sharing with her how she felt after her second divorce (Sanford 2021: 183–6) or that a play they went to see together in Boston in 1986, Sanctuary: The Spirit of Harriet Tubman, which drew connections between Tubman's anti-slavery crusade and Central American refugees’ experiences, stirred many memories of her own life. Both occasions are good examples of white centering: Sanford was busy thinking about a woman she fancied and prompted Mary to shift from her life (and her perspective) to hers. In consequence, though these forms of exchange have the surface appearance of EE, they fail in practice to count as the kind of epistemic engagement that can be a remedy for TI broadly understood. Sanford also recalls how when they first met she failed to understand that upon arriving after a long journey to a new home Mary would have wanted privacy or that she would not feel free to say no to the daughter of her employer (Sanford 2021: 7).

Although Linker has in mind debates about social identity and social difference in which interlocutors strongly disagree with one another and feel upset and hurt, her understanding of intellectual empathy could shed light on what is required for EE in any testimonial exchange, especially given that interlocutors are always socially situated. In that respect, a conversation does not need to be about social identity and social difference for social identity and social difference to play a big role in its development and its outcome. To give you an example, one of the leitmotifs of her book, namely that all of us enact social beliefs that could potentially anger or hurt others, could be reformulated in the following way: all of us hold social beliefs that could potentially discriminate against specific others. This formulation is consistent with her other idea that “we are all a work in progress” (Linker 2015: 158), as everyone is prone to get something wrong when it comes to beliefs about individuals and social groups.

Likewise, though Linker (2015: 131–3, 180) stays within the framework of virtue theory, the truth is that the skills she identifies under the umbrella of intellectual empathy also make sense from the perspective of EE as a practice. We could well be trained from early childhood into testimonial routines that make us develop the five skills that she identifies: (i) understanding the invisibility of privilege, (ii) knowing that social identity is intersectional, (iii) using the model of cooperative thinking, (iv) applying the principle of conditional trust, and (v) recognizing our mutual vulnerability (Linker 2015: 14).

Let us return to Sanford's recollections to understand why we need skills training. Essentially, Sanford does not awaken to Mary's world, fears and concerns, out of being virtuous, though virtues play a role in her realization of black rights, white wrongs, and social privilege. What ultimately prompted her to face her unexamined biases around class and race was her work as a women's health activist (Sanford 2021: 130). Her exposure to feminist issues and black feminisms in particular slowly provided her with the relevant knowledge and skills. For example, she describes in great detail throughout her book how African American literature allowed her to perceive the world around her in new ways; for instance, learning how it appeared to Mary. She calls that kind of reading ‘restorative’:

As I began to read literature by Black women and men more widely and with more purpose, I found myself adapting the traditional spiritual practice of ‘devotional reading’. My devotional reading for the rest of my life would feature, not the Bible and other more traditionally religious texts, but writings that increased my understanding of Black Americans’ history and culture, and also of white racism, white privilege, white supremacy. Over time, this devotional reading evolved into what I have come to call ‘restorative reading’. (Sanford 2021: 131)

Sanford understands restorative reading as “an ethical practice”, which allows her as a white woman “to access crucial information without burdening Black colleagues or friends” with her “blunt deficits”, and “to become more responsible and accountable in the real world” (Sanford 2021: 131). Other kind of exposure might also be restorative. Sanford calls attention to Mary's remark on the influence that watching Sanctuary could have had on the Caucasian audience surrounding her: “there I was with a whole roomful of Caucasians who were seeing for the first time what I was living. … I came later to understand that maybe some were really there for learning” (Sanford 2021: 189).

Let us return to Linker's list of skills. On my view, it is in cooperative thinking that all the skills that she distinguishes converge. She understands cooperative thinking as “thinking and reasoning cooperatively about social identity and difference”, so that we have access to “the relevant feelings, experiences, and data that are all necessary for understanding the oppressive aspects of social identity”, which “means hearing about how each of us experiences social systems and social categories”, as well as acknowledging structural pressures and “reducing their impact on our dialogues and discussions” (Linker 2015: 96). We need to understand one another as collaborators working together for the amelioration of collective pools of epistemic materials, which entails uncovering and understanding our social beliefs (Linker 2015: 97).

One of our first steps to achieve that goal needs to be gaining awareness that “most of us are susceptible to problematic stereotypes and generalizations” (Linker 2015: 100). We need to be trained to observe that all of us “hold social beliefs that could potentially anger or hurt others”, that “we all get something wrong when it comes to beliefs about individuals and social groups”, that “some of this is the result of ignorance of privilege, and some of it is just ignorance and a basic lack of experience and information”. In other words, we need to be trained to view each other “as collectively imperfect” (Linker 2015: 101).

Sanford exposes us to white people's imperfections by drawing our attention to her own attitudes in her relationship with her friend Mary. She recalls an occasion the summer they started spending more time together in which she hurt Mary's feelings by thoughtlessly saying to her that her mother was worried because she was not socializing enough, as if “socializing” did not include their time together (Sanford 2021: 145–6). Employing one of the concepts that she learnt from black activism, Sanford interprets this instance of negligence on her part as a microaggression. Along the lines of Linker, she claims that “we need to understand, in all humility, that we commit microaggressions all the time” (Sanford 2021: 146).

Sanford also shares with us a revelatory memory regarding the invisibility of social privilege. Shocked by her reading of civil rights activist Anne Moody's (1968) Coming of Age in Mississippi, she notices for the first time the whiteness of her own skin and that everyone else around her in the ferry to Nantucket is white as well (Sanford 2021: 136). She observes how her realization at the age of 37 of her own whiteness contrasts with seven-year-old Moody's awakening to the whiteness of her neighbors as her mother did not allow her to follow them into the “whites only” section at the movies. This is an example of how social privilege can be invisible for those who enjoy it. As Sanford explains, when white people are the norm, “affluent white people can reach the age of 37, or 73, or 100, without noticing their own whiteness” (Sanford 2021: 137). Reflecting on her realization of her white privilege, Sanford makes use of Peggy McIntosh's (2019) 1989 notion of “the invisible knapsack”: “an invisible package of unearned assets” which one can count on cashing each day, but about which one is “meant to remain oblivious”, so that one remains ignorant of the invisible system of dominance enacted by their own social group (Sanford 2021: 138).

We also learn from Linker that the fact that “we share the common experience of having been brought up by adults with biased and stereotypical beliefs”, “[t]hough the biases and stereotypes maybe about different groups”, “can often create a sense of common ground” (Linker 2015: 154). Sanford makes an impressive job at communicating many of the biased and stereotypical beliefs her parents inculcated in her and that her case was not exceptional, as all white girls in wealthy families were brought up similarly. However, she does not dwell upon the biased and stereotypical beliefs that might have given shape to Mary's childhood. Let me introduce you to another example in order to shed light on how knowing that all of us are vulnerable to biased and stereotypical beliefs can create a sense of common ground.

Norman Jewison's 1967 American mystery drama film In the Heat of the Night, based on African American John Ball's novel of the same name, tells the story of a black top homicide inspector from Philadelphia, Virgil Tibbs, involved in a murder investigation in a small town in Mississippi (Silliphant 1966). Late on Tuesday night, police officer Sam Wood discovers wealthy industrialist Phillip Colbert lying murdered on the street. Later that night, Wood finds Tibbs at the train station where he is waiting for a train as he is in the middle of his journey to visit his mother in Memphis. Without letting Tibbs speak, Wood arrests him. Once in the police station, police chief Gillespie accuses him of murder and robbery before learning that he is talking to a higher rank fellow police officer. As it happens, Tibbs and Gillespie end up collaborating to incarcerate the real murderer. However, their collaboration is neither straightforward nor ideal. Both before and after their collaboration starts, Tibbs is subjected to many instances of epistemic injustices owing to white racism. In fact, the fact that Wood arrests him without a single question could be understood as an example of pre-emptive testimonial injustice. Likewise, right at the beginning of the film, we witness a clear example of Fricker's original formulation of TI: Gillespie does not believe Tibbs when he says that he was waiting for a train to visit his mother.

As regards the example we are in search for, Tibbs and Gillespie visit Endicott, a bigoted cotton plantation owner who does have a motive to have killed Colbert. Endicott feels insulted for being questioned and slaps Tibbs, who delivers a harder slap back to him. Tibbs and Gillespie leave. Tibbs infuriatedly claims that he is determined to get Endicott for the murder and that he only needs two more days to bring Endicott right off his hill, which confirms Gillespie that Tibbs is capable of bias because of what Endicott represents in political terms. Gillespie, drawing attention to Tibbs’ biased attitudes toward Endicott, says to him: “You're just like the rest of us, ain't ya?” The viewer can tell from Tibbs’ face that he realizes that there is something in what Gillespie is saying. Later on, Tibbs acknowledges that he was so hung up trying to capture Endicott for personal reasons that he almost missed the truth. Despite the many imperfections of their cooperation, both Tibbs and Gillespie are aware of their own as well as of each other's limitations and point them out to each other. In the process, the amount of trust that they are able to extend to each other increases.

To help us deal with one another's limitations, Linker reminds us that all of us have an “uncle Moe” in our family: a person dear to us who openly expresses biased, stereotypical beliefs, with whom we might strongly disagree despite our love and care for them. Linker suggests that when someone expresses a biased, stereotypical belief, we engage with them as if they were our uncle Moe. We could extend this to any epistemic interaction. By engaging with our epistemic counterpart instead of writing them off as being unworthy of respect or of epistemic attention, we might find out that they were lacking a piece of information that might change their views or learn that we had interpreted them incorrectly. Besides being in a better position to share or gather relevant evidence, by EE we also gain “a better understanding of the speaker's perspective” (Linker 2015: 161). As far as EE as a remedy for TI broadly understood is concerned, as already mentioned, we would also gain a better understanding of their epistemic agency, as we would have more opportunities to evaluate their actual epistemic capacities.

Linker argues that we need to create a reasonable amount of trust and “begin with the presumption that people are trustworthy” and that we do have an obligation to understand and make sense of their views (Linker 2015: 156). Thinking of one's uncle Moe could be one of our strategies to foster the kind of trust that is needed for us to reason in cooperation, a “base level of mutual recognition for each other's vulnerability and humanity in the face of social pressures” (Linker 2015: 96). Linker puts forward ‘the principle of conditional trust’ (Linker 2015: 157) according to which we should trust and engage with one another provided that all parts (i) remain open to relevant counterevidence that the other person might provide, even if it challenges their beliefs, and (ii) enact reciprocal reassuring attitudes, such as acknowledging similar mistakes and communicating understanding if the other person retracts their claims and reassuring the other person that we won't distrust them if they raise the possibility that we are biased. Accordingly, we could say that one of the reasons why Tibbs and Gillespie end up trusting each other is because they allow the possibility that the other person suggests that they are biased, as discussed earlier. Among Mary's reassuring attitudes toward her friend Wendy, let me draw your attention to the occasion in which Mary requested Sanford that they continued a conversation they had started about race, warning her away from her self-centered awkwardness around the issue (Sanford, 2021: 150). We shall see later on that both epistemic interactions meet at least at some point the first condition.

In order to further clarify the kind of trust that we need to develop, I would like to make use here of my depiction of epistemic equality trust (Carmona 2021a: 586–7): a minimum of epistemic competence and willingness that, in the absence of counterevidence, we ought to presuppose one another on the basis of the epistemic potential of every human being. To avoid TV, the speaker's obligation, in the absence of counterevidence, is to accord the hearer a minimum level of competence and willingness. Likewise, in the absence of counterevidence, hearers usually accord speakers a minimum level of competence and willingness for the mere fact of being human beings. This kind of trust, which is rooted in a relational conception of human equality, is what makes testimonial exchanges function smoothly among strangers and might be required for avoiding TI as well as other forms of epistemic injustice. Accordingly, in the absence of counterevidence, EE would generally not only entail assuming epistemic competence on the side of our epistemic counterpart, but also presupposing a willingness on their part to participate in the epistemic exchange in question. I say ‘generally’ because oppressive power relationships, as we shall see in section 6.3, might make EE untenable.

To foster and convey this willingness, we can train ourselves to employ benevolent ways of arguing; for instance, we should train ourselves to find opportunities to remind one another that certain conversations are difficult or risky, as well as to develop reassuring attitudes. This is intimately related to the acknowledgment of our mutual vulnerability, which “means that we go into the conversation ready to be affected and willing to be transformed” (Linker 2015: 165). In this regard, acknowledging one's vulnerability requires courage and strength (Linker 2015: 165). To make the most of our openness to self-correction and transformation, we also need metacognitive skills training. We would need to participate in routines that reinforce our ability to think about what we think and sensitize us “to our experiences of cognitive dissonance, or the state of having contradictory beliefs as well as beliefs that do not seem to match up with reality” (Linker 2015: 96).

Likewise, our sense of common ground is likely to be reinforced if we have an intersectional conception of identity. By being sensitive to the variety of grey shades that there exists between what we frequently understand as opposites, we might build bridges with our fellow human beings. As Linker explains, this might in itself help relax our blame and guilt filters, which might lead to a variety of connections and coalitions among diverse social groups; for instance, we might realize that even if not owing to the same aspect of our identity, we share the experience of marginalization (Linker 2015: 70–6). Accordingly, besides training ourselves to identify bias, we need to learn to distinguish the different layers of our identities.

Sanford (2021: 126–7) explains that it was not until she had learnt to scan the world intersectionally, that she realized the specific struggles that black women had to bear and was able to detect “the socially constructed walls” that separated Mary from her. Likewise, intersectionality was a door into Mary's experiences and concerns; for instance, into what her first marriage meant for her, something that Sanford was unable to grasp before being transformed by feminism and her restorative reading (Sanford 2021: 84–5).

Similarly, we can learn much from In the Heat of the Night as regards how positive for EE is to have an intersectional account of social identity. The film builds a contrast between black police inspector Tibbs, the country's foremost homicide expert, college-educated, well-dressed, refined manners, and with a healthy, elegant complexion, and poorly paid white police chief Gillespie, who works in a shabby office in a seedy town in which not even his subordinates respect him. Unlike Tibbs, who has succeeded professionally, Gillespie is not satisfied with his life and is troubled by his loneliness. When Gillespie learns about the complexity of Tibbs's identity, in particular that he is a fellow police inspector who earns much more than him, Gillespie starts trusting him epistemically. The movie is excellent at showing that Gillespie initially only respects Tibbs because of his manifest knowledge, using and objectifying him. Only later in the film does Gillespie show concern for Tibbs as a human being as well as interest in hearing about the details of his existence. The change in Gillespie's attitude can be understood as an example of the benefits of awareness of intersectionality. In Tibbs’ case, some aspects of his identity mark him as marginalized and others mark him as privileged. That complexity contrasts with (and disrupts) the stereotype with which Gillespie first approaches Tibbs. Not only does Tibbs’ complex identity motivate Gillespie toward trusting him epistemically, helping him to individuate Tibbs, but it also makes him notice the affinities they might have; for instance, that they both live alone and don't have their own family.

The truth is that I cannot think of a more enlightened depiction of the struggle that proper, authentic EE can sometimes involve than the cooperation between Tibbs and Gillespie in In the Heat of the Night. Though Gillespie, as we said, starts trusting Tibbs epistemically when he learns about his otherwise manifest expertise, he does change his attitude towards him after the arrest of a white man, Harvey Oberst – another example of intersectionality in the film, as he portrays himself as someone who has never had any luck in his life despite being a white heterosexual male. Suddenly, Gillespie, content with the idea that he can finally close the case, stops being interested in what Tibbs has to say and tells him that he is no longer needed; for instance, Gillespie does not want to learn about the results of the autopsy. Tibbs refuses to leave and provides Gillespie with evidence about Oberst's innocence. In a sequence of scenes, we can see how Gillespie painfully reconsiders his view. It is as if he were having a pitched battle against his prejudices which manifests bodily. In the end, Gillespie, convinced by Tibbs’ arguments, drops the murder charge and accuses Oberst of robbery instead. Something similar occurs later in the film when Gillespie arrests fellow police officer Wood for the murder. Tibbs has to work really hard and collect unquestionable evidence to convince him that Wood is innocent, but Gillespie engages epistemically with Tibbs and does ultimately change his mind. I think we could say that this is a realistic example of the fulfillment of one of the prerequisites for Linker's principle of conditional trust, namely that we remain open to counterevidence even if it challenges our beliefs.

What do we essentially learn from these examples? That EE is about exposing oneself to one another's epistemic aptness, that a minimum of (conditional) trust is needed for that, that we need to learn to identify bias as well as to distinguish the different layers of our identities, that we need relevant information about social privilege and realize our mutual vulnerability. In addition, we grasp from them that EE is so demanding that we need skills training, for the struggle of fictional chief police Gillespie might have been significantly less if he had been acquainted with the kinds of strategies and skills that Sanford learnt from the women's health movement as well as from black activism and her restorative reading, not to mention that something which is essential for social cohesion should not be left to the virtuosity of individuals.

5.3. Consequences for the concept of testimony

The employment of the strategies and skills of epistemic interaction mentioned in sections 5.1 and 5.2 ought to transform our conception of testimony. We need to stop conceiving our testimonial exchanges as single transactions in which a hearer is supposed to receive (and assess) epistemic materials from a speaker. By contrast, testimony ought to be understood (and practiced) as a product of the active epistemic engagement between epistemic agents who might interchange their roles, in such a way that speakers become hearers and vice versa. In consequence, EE casts light on the fact that fighting TI requires second-order change, in other words, a change in the way we conceive of testimony as well as a change in the ways in which we engage with each other when testimony is at stake.

Dotson's (2012) order-of-change approach differentiates between first-, second-, and third-order change. First-order change entails incremental modifications within an established framework. By contrast, second-order change is “located at the level of frameworks and structures themselves”: it is what Dotson (2012: 31) identifies, following Langton (2010: 463), as a conceptual revolution. While second-order change “involves altering a single set of shared hermeneutical resources”, third-order change calls for “the ability to shift hermeneutical resources, which requires fluency in differing hermeneutical resources” (Dotson 2012: 34). Dotson argues that TI is a first-order epistemic injustice because ameliorating TI does not challenge credibility. Rather, “addressing problems of unwarranted credibility deficits requires reform within the framework where credibility still confers authority” (Dotson 2012: 28).

If our concept of testimony is transformed by EE as suggested, our understanding of credibility will also be modified. Credibility ought not to be understood simply as a property of a potential epistemic interlocutor – one that we may need to judge, and if so we should judge as objectively as possible, but in which we are not actively involved. By contrast, credibility needs to be understood as a relation in which one party stands to another, and one that therefore implicates us as participants in a dialogue, and not merely as observers. Though EE could be understood as a question of incremental modifications within the framework of testimony – that is, first-order change, such a practice entails a conceptual revolution regarding our concept of testimony – namely, second-order change. Such an interaction between first- and second-order change responds to Dotson's (2012: 36) insight that first-, second-, and third-order change “cannot be genuinely separated except in theory”.

5.4. Plenty of room for the virtuous agent

EE has room for the virtuous agent. One does not need to understand individual psychological mechanisms as the essential point of intervention to accept that orienting our lives according to the right virtues can be positive to ameliorate epistemic injustice. Recall Sanford's emphasis that we need humility to be able to understand that “we commit microaggressions all the time” (Sanford 2021: 146). Likewise, we learn from Mary's 2020 phone remark to her friend Wendy that if Sanford had not been a curious person, she would remain ignorant of Mary's world even today (Sanford 2021: 305). That said, this necessarily entails being aware of the limitations and worries of a virtue approach (Alfano 2015; Sherman 2016a). For instance, realizing that one is likely to remain unaware of one's blind spots. Or the awareness that “a mistaken idea about what is virtuous could always lead someone astray” (Sherman 2016a: 239). Accordingly, one also needs education in this respect.

It is important that the remedies we propose to tackle epistemic injustice are as broad and chameleonic as possible, as we need a particularist approach that finds solutions for real, idiosyncratic injustices (Anderson 2011; Medina 2013: 11–12). Accordingly, we ought to make use of all the tools at our disposal (Davidson and Kelly 2015). In this regard, there are two sides to EE as a collaborativist solution:

1. (i) Individual epistemic agents are supposed to collaborate with one another for testimony to be successful, and

2. (ii) EE can be coordinated with other remedies to fight TI. For instance, EE can be enacted by virtuous agents.

I agree with the insight that there is no such a thing as a virtue of testimonial justice (Medina 2013: 23; Sherman 2016a). Which virtues are helpful in a specific testimonial exchange will be context-dependent. EE is broad enough as a practice to incorporate manifold virtues. Medina (2013) calls attention to epistemic humility, curiosity/diligence and open-mindedness. Besides humility, Sherman (2016a) favors moral justice, epistemic responsibility and vigilance. The core virtues of care (Engster 2005: 54–5) would also be helpful in our testimonial interactions: attentiveness, responsiveness and respect. There is not a closed list of the virtues that might be helpful in the context of TI broadly understood.

Unlike other proposals (Fricker 2007; Medina 2013), despite making room for virtue, EE is not rooted in virtue theory. Besides, the notion of virtue that I have in mind is Humean, instead of the Aristotelian model that dominates most virtue theory (Davidson and Kelly 2015: 19; Merritt 2000; Sherman 2016b).⁴ On my view, battling against TI broadly understood requires an epistemic subject that while maybe acting from the virtues is not committed to any strong ideal of motivational self-sufficiency of character. I say ‘maybe’ because though virtues might help while EE, they might not be necessary, as the strategies entailed by EE might suffice to avoid instances of TI, especially when they become second nature. In this regard, EE has a direct impact on our testimonial practices and thus contributes to give democratic shape to our epistemic ecology.

Attention has been drawn to promising ecological approaches that focus on changes in the epistemic environment to fight TI (Alfano 2015; Davidson and Kelly 2015; Sherman 2016b; Washington 2016). Shifting attention from individual minds, EE exerts ecological control on our judgments. The practice of EE is an environment-engineering intervention aimed at fixing how we engage epistemically with one another in our testimonial exchanges. Instead of focusing on self-correction and self-alertness, EE strives for transforming the dynamics of testimony in such a way that our testimonial practices are eventually embedded in a democratic epistemic ecology that has a direct impact on our values.

6. Some of the Benefits and Limitations of EE

For reasons of space, the list of benefits and limitations provided here does not aim to be exhaustive.

6.1. Fostering epistemic democracy

One essential benefit of EE is that, by promoting epistemic interaction, it might help current societies to embrace the ethos of epistemic democracy. We learn from Medina (2013: 3–13) that interaction is essential to epistemic democracy. His notion of interaction

calls mainly for communication and cooperation, that is, for developing the ability to share social spaces and practices without ignoring each other, but fully taking into account each other's different experiences, interests and aspirations. (Medina 2013: 9)

EE fulfills Medina's Imperative of Epistemic Interaction according to which “democratic sensibilities require free and equal epistemic interaction among the heterogeneous groups that are part of society”, which “calls for the development of communicative and reactive habits that operationalize our responsiveness to diverse and multiple others” (Medina 2013: 9). To be explicit, EE calls for appropriate attentiveness and responsiveness to one another's epistemic aptness in our testimonial practices. By ‘democratic sensibilities’, Medina (2013: 9) has in mind “cognitive-affective attitudes that facilitate and promote the capacity to relate, to listen, to feel concerned, and to care for the interest and aspirations of others”. As we saw in sections 3, 4, 5.1 and 5.2, the practice of EE addresses the cognitive, affective, and communicative attitudes one needs to develop such a democratic (epistemic) sensibility, enabling those with enough training “to share spaces responsively and to engage in joint activities” (Medina 2013: 9).

Epistemic interaction in Medina's (2013: 4) proposal is not oriented toward consensus but rather toward resistance, which he understands as the centerpiece of a democratic culture. Observe that EE also fulfills the two basic principles of Medina's (2013: 50) epistemology of resistance:

1. (i) Principle of acknowledgment and engagement: “all cognitive forces we encounter must be acknowledged and, insofar as it becomes possible, they must be in some way engaged”.

2. (ii) Principle of epistemic equilibrium: “searching for equilibrium in the interplay of cognitive forces, without some forces overpowering others, without some cognitive influences becoming unchecked and unbalanced”.

By definition, EE entails engagement with one's epistemic counterpart. It is by engaging with one's epistemic counterpart that one acknowledges them as a knower. Concerning epistemic equilibrium, EE as a practice aims at ameliorating epistemic injustice, working toward a society of epistemic equals, in which all perspectives are engaged.

Another fundamental benefit of EE is that its core strategies, individuation and contact, expose epistemic agents to one another's specific problems and hermeneutical devices. If Medina (2013) and Pohlhaus (2012) are right in thinking that hermeneutical injustice (HI) lies beneath testimonial injustice (TI), as I believe they are, the remedies we propose ought to have a positive effect on our hermeneutical awareness. Recall the discussion of TS in section 4. By EE, we are exposed to the point of view of other epistemic groups. Accordingly, we are likely to be more sensitive to their problems, their hermeneutical resources and their expressive manners. Consequently, owing to the insight that there are other ways of looking at the world, we are more likely to revise our perspective. As a result, we might be able to fight HI as well as the meta-insensitivity characteristic of situated ignorance. If HI is at the core of TI, meliorating the former is likely to have a positive effect on TI. EE might also help preventing contributory injustice (Dotson 2012), as such exposure to the epistemic other's hermeneutical devices might encourage epistemic agents to shift hermeneutical resources. In Dotson's (2012: 31–2) terms, EE might thus encourage third-order change.

6.2. Less room for moral and epistemic bad luck

Fricker (2007: 103) depicts moral and epistemic bad luck as circumstantial – affecting what one does – and constitutive – affecting who one is. In her view, achieving TJ might be altogether impossible under certain circumstances. She puts forward that a certain culture at a given time might not provide its people with the critical reflective tools required to achieve the virtue. Such an idea allows Fricker to introduce her concept of “non-culpable TI”, an injustice for which the perpetrator cannot be blamed.

Fricker interprets as an example of moral and epistemic bad luck the case of Greenleaf in Anthony Minghella's screenplay of The Talented Mr. Ripley. Fricker argues that the lack of concepts in the social imaginary available to Greenleaf at that point of time constituted him as someone who could not deliberate in favor of women against gender prejudice. From Fricker's perspective, Greenleaf did not have access to the reasons to deliberate in favor of women. Accordingly, he could not be blamed for failing to do so. Therefore, Fricker (2007: 100) concludes that Greenleaf was not culpable.

EE allows less room than Fricker does for what she calls “moral and epistemic bad luck”. As already argued (cf. Coady 2012; Dotson 2012), this is necessary. By EE, one's epistemic point of view might benefit from friction between prejudice and evidence obtained in the interaction. The more interaction with one's epistemic counterpart, the more evidence we might obtain to fight prejudice, even in cases in which historical circumstances make it particularly hard. Consider again the (lack of) interaction between Greenleaf and Marge. As I pointed out earlier, if Greenleaf had engaged with her in conversation to find out whether she actually was aware of his son's sordid habits, he would have realized that she did not need protection from such epistemic materials. In other words, Marge's knowledge of the tawdry details of her fiancé's life would have worked as a piece of evidence against gender prejudice.

Fricker leaves the virtuous hearer on their own. In this regard, her notion of TJ is implicitly committed to the individualist perspective on agency and reasoning in the discussions of strategies to ameliorate the effects of prejudice (Washington 2016). By contrast, EE acknowledges the role of the one at the receiving end of the epistemic injustice in reconditioning the perpetrator's testimonial sensibility, as engagement with one's counterpart's epistemic agency can help to ameliorate TI broadly understood. (Observe that I don't mean to say that one's epistemic counterpart ought to make the effort to educate us. I will return to this point in the next section. What I mean is that exposure to one's epistemic counterpart's epistemic aptness might, by itself, ameliorate epistemic injustice in our testimonial exchanges.) Even if Fricker takes the social atmosphere into account – for instance, by referring to the social imaginary – speaker and hearer remain as remote from each other as isolated islands. The nature of identity prejudice might indeed be separating them. However, if our correctives for epistemic injustice in our testimonial exchanges were understood – and practiced – in a more dialogical way, another day would dawn. In this regard, EE is a collaborativist solution and a drastic shift away from individualism.

6.3. Limitations owing to the impact of oppressive power relationships

I agree with Bowell (2021: 103) that for our considerations of engagement to be responsible we need to take power asymmetries into account. Epistemic agents are always socially situated in relation to one another. The epistemic agents of our core examples, Mary Norman and Wendy Sanford in These Walls Between Us and Gillespie and Tibbs in In the Heat of the Night, are differentially empowered. In addition, every epistemic exchange is socially located. We need to take into consideration that EE takes place within the frameworks of structural epistemic injustice. While there is insufficient space in this paper to tackle this question, I feel the need to acknowledge it as an issue both for my own account of EE and for Fricker's (2007) work more generally.

The culprit-based epistemic injustices identified in this paper are also fueled with structural injustice. Dotson (2012) and Medina (2013), inter alia, have shown that it is not possible to separate testimonial injustice from structural hermeneutical injustice. The reason why Sanford finds it hard to put herself in Mary's perspective is that the hegemonic collective hermeneutical resources that she finds at her disposal in a society in which white is the norm do not include those derived from black people's specific experiences and concerns. As a result, in order to be able to shift perspectives, Sanford trains herself in their marginalized hermeneutical resources and experiences by her restorative reading. It is revealing that not until she learns from activists the concept of ‘microaggression’ did she understand fully some of the occasions in which she had hurt her friend Mary (Sanford 2021: 146).

Sanford shares with her readers that she often feared that Mary would retreat and decide not to engage with her any further; for instance, after watching Sanctuary, as the play made white oppression tangible on the stage, or in the many occasions in which she failed to put herself in her shoes (cf. Sanford 2021: 124, 145–6, 185). In that respect, she constantly refers to Mary's generosity, which makes me wonder whether it is reasonable or, for that matter, just, to expect the marginalized to engage readily with those non-marginalized or with the less marginalized. Should the marginalized provide epistemic resources to those situated away from the margins in order to facilitate their understanding of aspects of social reality that pass unnoticed from their perspective? Should they do it on every occasion or are there times in which they should not? Recall that one of the reasons why Sanford becomes immersed in her restorative reading is because she wants to avoid “burdening Black colleagues or friends” with her “blunt deficits” (Sanford 2021: 131). Should we follow Sanford's example? Should those with greater power be willing to do more of the epistemic heavy lifting in a way that is not wholly reliant on the experiences and testimonies of the less empowered?

Pohlhaus (2011) argues that when comprehending certain positions or lines of thought unfairly undermines the agency of a socially situated knower it is wrong to ask for their understanding. She draws from Maria Lugones’ work to clarify her point: marginalized knowers shouldn't be asked

to inhabit worlds that oppressively constrain their agency, including their epistemic agency, so that it is impossible to fill the request of understanding without simultaneously foreclosing the ability to demonstrate the harm perpetuated by the world that sustains the understanding. (Pohlhaus, 2011: 231–2)

We learn from Pohlhaus that there are certain calls for neutrality and for “understanding all sides” that are anything but neutral. In fact, they make marginalized positions invisible because “demonstrating the harm that the requested understanding does can only be done from worlds that actively resist the sense of the world one has been implicitly asked to inhabit” (Pohlhaus 2011: 232). Let us go back to In the Heat of the Night. Everyone with the exception of Colbert's wife shows racist attitudes towards Tibbs. Would we have the right to ask Tibbs to understand and engage with their point of view? I agree with Pohlhaus that we should not, as Tibbs would have to participate in a set of meanings that constitute him in ways that undermine his epistemic agency and his very humanity. For instance, Endicott's slap is intimately connected with his view that black people are like orchids that need care, feeding, and cultivation. Likewise, if Tibbs were asked to understand the reasons why he was arrested by Sam Wood or why Chief police Gillespie did not believe his word that he was waiting for a train to visit his mother, he would be forced to consider black people as potential criminals and liars. However, though Tibbs does not have an understanding attitude toward them, he does grasp how they behave and how their individual behavior is embedded in structural white dominance. This understanding of his, which he needs in order to survive in the world around him, in which whiteness is the norm, is neither empathetic nor tolerant.

Bowell explains that exercising tolerance in cases in which “the marginalized are asked to engage from a position of vulnerability, specifically where they are asked to attempt to understand the standpoint of the dominant” (Bowell 2021: 100) is “imprudent because of the harm it does to oneself, to those in a similar position, or those who are part of one's community” (Bowell 2021: 102). In fact, she goes so far as to argue that in such cases the right thing to do “in order to care for oneself and one's community” is “to withdraw from critical engagement” (Bowell 2021: 102).

As Bowell (2021: 103) suggests, in situations of asymmetrical power,

the enactment of argumentative vices is structural, whereas the expectation of empathic critical engagement – demands for understanding – come to bear on individuals. Power asymmetries thus take on an additional dimension whereby the marginalized individual is responding to requests that take place within the asymmetric political, social, cultural, and bureaucratic structures that create and perpetuate those asymmetries.

Consequently, there is another layer to why we cannot ask Tibbs to have an understanding attitude toward someone like Endicott: the request that Tibbs engages epistemically with Endicott is rooted in a context that actually reinforces Tibbs marginalization. Moreover, returning to our analysis of the individual level, Endicott, unlike Gillespie, does not meet the prerequisites of Linker's principle of conditional trust. The way Endicott speaks of black people makes us realize as viewers that he is neither willing to change his views nor to consider that he is biased, as it is obvious that he wants to hang on to a world that starts falling apart.

When I claimed in the Introduction that a fundamental feature of this proposal is the vindication of the role of the one at the receiving end of the epistemic injustice in reconditioning the perpetrator-to-be's testimonial sensibility, I did not mean that marginalized knowers ought to EE with dominant knowers, nor that the former ought to make the effort of educating the latter. By contrast, the idea that I want to communicate is that non-marginalized knowers’ sensitivity and attitudes might be reconditioned by training themselves in EE and being exposed to marginalized knowers’ epistemic agency. To give you an example, Sanford tells us about one occasion in which she realized her social privilege as a direct consequence of EE with Mary's perspective (Sanford 2021: 195). Prompted by Sanford's insistence to take a holiday the following summer, Mary asked Sanford whether she had ever tried to say no to her mother, to which Sanford thoughtlessly responded affirmatively, but as she spoke, the privilege of being able to say no to her mother dawned on her. Consequently, what initially would only have had the surface appearance of EE without being EE in actual terms became an instance of authentic, transformative EE.

On my view, the question whether marginalized knowers are supposed to engage with non-marginalized knowers is intimately related to Dotson's (2011) concept of testimonial smothering (TS). As explained in section 4.3, TS takes places when a marginalized speaker truncates their testimony because they expect to be misunderstood by an audience who fails to show testimonial competence. Dotson argues that not only is the marginalized knower not required from the point of view of epistemic equality to share the epistemic materials that they have and that the non-marginalized ignore, but also that they are the one at the receiving end of the injustice though remaining silent. By contrast, non-marginalized audiences, upon listening to unsafe and risky content coming from marginalized knowers, need to meet extra demands; for instance, indicating testimonial competence (Dotson 2011: 250–1). Similarly, as I have explained elsewhere, there are extra demands coming from marginalized audiences that a speaker needs to address regarding risky epistemic materials, in such a way that “for speakers not to exercise epistemic violence on hearers regarding risky content, speakers ought to exercise testimonial competence, which essentially amounts to acknowledging the epistemic point of view of audiences” and sharing epistemic materials accordingly (Carmona 2021a: 589). Let these examples serve as an affirmative answer to the question whether those with great power should be willing to do more of the epistemic heavy lifting in a way that is not fully reliant on the experiences and testimonies of the less empowered. May they also be understood as a negative reply to the question whether it is the responsibility of marginalized knowers to engage readily with the more powerful and educate them. However, if marginalized knowers feel safe enough, as Mary did in respect of Sanford when their connection began to resemble a friendship, it is more than desirable that they do EE with non-marginalized knowers, as marginalized knowers have much to teach the privileged. That said, for marginalized knowers to engage with non-marginalized knowers the latter should meet the two prerequisites of Linker's principle of conditional trust, as Sanford did.

Let me close this section by saying that the fact that we might not be able to make a given engagement an equal one should not prevent us from EE whenever safe and possible. We should not aim at creating perfect public discourse, as it would not be realistic and could paralyze us. Rather, we want to foster “a more improved discourse” (Linker 2015: 156). Hopefully, developing the core strategies and skills mentioned in this paper would help to build sufficient trust to enable a fair enough EE in those cases in which marginalized knowers’ epistemic agency and humanity are not under risk.

7. Final Remarks

TV, along with other kinds of TI identified in the last decade, demands the revision of TJ as the remedy for TI broadly understood. I propose that epistemic agents ought to be trained in the practice of actively engaging epistemically with one another in our testimonial practices. Besides ameliorating cognitive attitudes, EE would improve affective and communicative attitudes, as we are more likely to give and receive epistemic materials with the right empathetic attitude and will have more exposure to the hermeneutical resources of our epistemic others.

Though it might sometimes be impossible or even undesirable (for instance, in situations in which oppression makes EE unsafe and risky for marginalized knowers) to further interact with one's epistemic counterpart in a testimonial exchange, the experience of EE in certain cases could have a positive impact on how we generally approach others in our everyday epistemic exchanges, even in the most hurried and abrupt cases. In this regard, EE can contribute to the development of the baseline level of epistemic trust that should be extended toward every human being in an epistemic democracy. In other words, it could smooth the way for a society of epistemic equals by fostering a more improved public discourse.⁵