Simulating experiences: unjust credibility deficits without identity prejudices

ABSTRACT

This article focuses on unjust credibility deficits in cases of testimony about emotional reactions towards acts of oppression. It argues that the injustice in these cases is not rooted in the hearer’s identity prejudices against the speaker, but the hearer's problematic way of dealing with his simulation of being in the speaker's situation. The simulation is in itself not morally problematic. However, I focus on a case where the hearer either recklessly or negligently fails to consider knowledge about the differences in people’s emotional reactions and comes to believe that his simulated reaction is the only possible reaction, thus believing that the speaker does not report her emotions truthfully.In this case, the ‘epistemic poison’ results from the recklessness or negligence not necessarily from identity prejudice against the speaker. This credibility deficit is unjust, yet its persistent and systematic nature stems from the fact that a hearer who is not from the same marginalized or oppressed group as the speaker will simulate different emotional reactions to those reported by the speaker. Hence, the unjust credibility deficit can be connected to other forms of social injustices suffered by the speaker.

KEYWORDS:

• Empathy
• simulation
• epistemic injustice
• testimonial injustice
• identity prejudice
• interpersonal understanding

1. Introduction

In August 2023, Spain won the football World Cup. When the team collected their gold medals, the president of the Spanish football federation, Luis Rubiales, grabbed the forward Jennifer Hermoso by her head, pulled her toward him, and kissed her on the lips. She later claimed, spontaneously, that she ‘didn’t like it’. Despite her report that she didn’t like it, the questions of whether she really felt uncomfortable and whether the kiss was really unwanted later became matters for public debate.

The unwanted kiss was clearly a violation of a woman’s body and her autonomy and integrity. Nevertheless, a powerful man acted in a way which seemed natural and unproblematic to him. His comments about the incident suggest that he does not think or cannot even imagine that a woman in Jenni Hermoso’s position would not appreciate his behavior.

It is also a feature of this and many similar cases that the act of sexual harassment is not seen as such, either by the harasser or by other people discussing the case, despite the victim’s testimony that the act was not wanted, a threat or anything along these lines. When something like this happens, we can observe a second moral wrong. This moral wrong consists in the way the victim is treated in her role as testifier, as was observable in the debates about Hermoso’s comments on the kiss. Some commentators did not, it seems, believe Hermoso when she reported that the kiss was unwanted.

The aim of this paper is to look closer at these cases and to argue that those who speak out often suffer from unjust credibility deficits because the hearer fails to imagine or simulate the psychological reaction reported by the speaker.

To lay to groundwork for my argument, I will first introduce the notion of testimonial injustice as formulated by Miranda Fricker (2007). It will become clear that certain forms of belief, namely those which Fricker calls identity prejudices, play a central role in the central cases of testimonial injustice.

In a second step, I will introduce the simulation theory of social cognition and its central claim that we understand others by means of mental simulation. I will distinguish between two ways of simulating: (a) simulating being in the other’s situation; and (b) simulating being the other in their situation. The content of both simulations can be either merely propositional or experiential.

With this work in hand, I will focus on testimonies about experiences that are relevantly similar to Hermoso’s report that she did not like the kiss. I will argue that we can find many cases where people merely propositionally simulate being in the other’s situation. I will focus on simulations which have an entirely different reaction as an outcome as claimed by the speaker. The simulator in this case comes to believe that the simulated reaction is the only possible reaction and therefore that the speaker is not telling the truth. I will assume that the credibility deficit in these cases harms or can harm the speaker to the same extent as in Fricker’s central cases. I will, however, argue, that we can find cases where identity prejudices influence neither the simulation process nor the way of dealing with it.

2. Testimonial injustice

Marginalized groups suffer from many different forms of injustice. In her 2007 book, Miranda Fricker focuses on a specific form of injustice which concerns members of marginalized groups in their roles as epistemic agents. She argues that, from the exclusion from education to the credibility deficit, we can find different varieties of epistemic injustice. What is special about this form of injustice is that it is not only a moral wrong, but is somehow epistemically wrong, too. Fricker’s book turned (analytic) philosophy’s attention to the topic of epistemic injustice and was the starting point for nuanced debate (for an overview, see Kidd and Medina 2017). Whilst previous debates in epistemology on testimony and knowledge transfer focused on how the hearer depends on the speaker and her credibility or trustworthiness, debates on epistemic injustice highlight the speaker’s dependence on the audience (Dotson 2011; Fricker 2007; Hornsby 1995). In the following, I will focus on what Fricker calls testimonial injustice, an injustice which occurs when a speaker does not get the credibility she deserves.

Sometimes people have good reasons not to trust; to disbelieve or even ignore what another person says. If I know that Mildred is a notorious liar, I have good reason not to trust her. If I know that Malachy has very right-wing political positions, I will not believe in his judgments and – at least under certain circumstances – ignore his arguments. Suffering from a credibility deficit is thus not always a case of testimonial injustice. The reason for the credibility deficit, i.e. the reason why the hearer distrusts or ignores the speaker, has to be in some sense unjust. Fricker is mainly interested in cases where a speaker suffers from credibility deficit because of an identity prejudice held by her hearer. An identity prejudice is a prejudice which concerns a person’s identity as a member of a social group. A prejudice about women as women is an example of identity prejudice. Very often, the relevant prejudices concern cognitive capacities or trustworthiness.

Fricker argues that, in the central cases of testimonial injustice, speakers suffer from a credibility deficit on the part of the hearer as a result of that hearer’s identity prejudice. An example would be when a hearer does not believe a person because she is a woman. If testimonial injustice is connected with other types of injustice via a common prejudice, Fricker calls it ‘systematic testimonial injustice’:

Systematic testimonial injustices, then, are produced not by prejudice simpliciter, but specifically by those prejudices that “track” the subject through different dimensions of social activity—economic, educational, professional, sexual, legal, political, religious, and so on. Being subject to a tracker prejudice renders one susceptible not only to testimonial injustice but to a gamut of different injustices, and so when such a prejudice generates a testimonial injustice, that injustice is systematically connected with other kinds of actual or potential injustice. (Fricker 2007, 27)

A central case of testimonial injustice is identity-prejudicial credibility deficit. In her book, Fricker describes a fictional example of a central case of testimonial injustice taken from Anthony Minghella’s screenplay of the 1955 Patricia Highsmith novel, The Talented Mr. Ripley. In the film, the female character, Marge, believes that Tom Ripley has murdered Dickie. When she tells Dickie’s father, Herbert Greenleaf, about this, he does not believe her and says: ‘Marge, there’s female intuition, and then there are facts’ (Minghella 2000, 130). Marge becomes marginalized in her concern. Mr. Greenleaf does not agree with her and, more importantly, does not even consider what Marge says. For him, Marge’s worry is not even worth thinking about. The quotation above suggests that this stance toward Marge is the result of his prejudices against women. As a woman, Marge is considered unable of thinking clearly about the situation and unconcerned with facts. Therefore, the reason why Greenleaf does not listen to Marge is his identity prejudices against women. Because of the unjust nature of these prejudices, Greenleaf’s behavior is morally wrong. It is also epistemically wrong because, for epistemic reasons, it would have been better to believe Marge since she is right. Believing her would probably have helped to solve the murder case and arrest Tom Ripley.

Many authors have analyzed how a credibility deficit constitutes or leads to serious harms (for example Collins 2000; Dotson 2011; Langton 1993; Tuana 2006). Fricker distinguishes a primary from a secondary aspect of harm (Fricker 2007, 44ff.). The primary harm is a form of essential harm. The speaker is wronged in her capacity as a knower. She is prevented from giving her knowledge to others and this is, as Fricker argues, an essential human value. Fricker goes on to argue that.

When someone suffers a testimonial injustice, they are degraded qua knower, and they are symbolically degraded qua human. In all cases of testimonial injustice, what the person suffers from is not simply the epistemic wrong in itself, but also the meaning of being treated like that. Such a dehumanizing meaning, especially if it is expressed before others, may make for a profound humiliation, even in circumstances where the injustice is in other respects fairly minor. But in those cases of testimonial injustice where the driving prejudicial stereotype explicitly involves the idea that the social type in question is humanly lesser (think of the sort of racism heaped upon Tom Robinson—“all Negroes lie”), the dimension of degradation qua human being is not simply symbolic; rather, it is a literal part of the core epistemic insult. (Fricker 2007, 44–45)

Beside the essential harm and its symbolic dimension, Fricker also highlights the practical and epistemic dimensions of harm as secondary aspects. Here we can think of the negative consequences of testimonial injustice in many different respects and in many different areas of life, such as the negative consequences in one’s career if one is confronted with the power of the police or in court but also exclusion from education.

In the following, I will focus on a particular form of testimony; namely, testimony about a speaker’s own experiences. I will analyze a case where the speaker suffers from an unjust credibility deficit. I will assume that in this case, the unjust credibility deficit has the same dimensions of harm as Fricker describes in her central cases. I argue, however, that in these cases we can assume that the reason for this unjust credibility deficit is not identity prejudice against the speaker on the hearer’s side. The reason is instead a reckless or negligent way of dealing with a simulation project. Finally, I will argue, similarly to Fricker in her central case of testimonial injustice, that the unjust credibility deficit is persistent and systematic.

The next section presents different simulation projects. These distinctions will help to identify the simulation project in the case I will analyze in section 4.

3. Simulation projects: four distinctions

Many recent accounts of social cognition and empathy hold the view that the process of mental state recognition includes a process of imagining or simulating the target person’s mental situation (Bailey 2022; Goldman 2006; Stueber 2006; for an overview Maibom 2017). Simulation theory was initially developed as an alternative to so-called theory theory. In broad terms, theory claims that we come to know which mental state another person is in by means of inferences based on a folk psychological theory of mind. Simulation theorists criticize this account for being far too intellectualist and offer an alternative: instead of inferences based on a theory, we come to recognize another’s state by means of simulation based on our own psychology. To take a rather unproblematic example: if I want to know why someone goes to the fridge to get a beer, I can use my own psychology as a model of simulation to arrive at the thought that I would do this too if I had the desire to drink a beer and the belief that there is beer in the fridge. What happens in this example is the following: the simulator puts herself in the situation of the target person. She simulates what she would think, feel, or desire in a specific situation (Stueber 2006; 2023).

In what follows, I will apply a distinction to simulation which is common in the philosophy of imagination; namely, the distinction between imaginings with merely propositional content and imaginings with experiential qualities. There is a debate about the question of whether purely propositional states are imaginings at all. Skeptics argue that seeing these states as imaginings expands the realm of imagination in an unhelpful way because states like assumptions and suppositions and many cases of counterfactual thinking would count as imaginings in this sense (Kind 2001; Langland-Hassan 2015; 2020). Others seem to accept this consequence and see it as supporting the important role of imagination in our mental life (Carruthers 2002; Currie and Ravenscroft 2002; Nichols 2006; Nichols and Stich 2000). In the following, I will use terms such as ‘propositional imagination’ and ‘propositional simulation’. However, it is not important for my arguments whether these states count as imaginative states or whether they must be classified in a different way.

In order to get the best explanation for the act of getting a beer, a merely propositional simulation project is sufficient. By a simulation or imagination project, I mean a simulation or imagination activity with a beginning and end point in time. It may or may not be started willfully. Propositionally simulating being in another’s situation means to imagine merely propositionally that I am in the target person’s situation. In the example given, this includes imagining that I have the belief that there is beer in the fridge and that I have the desire to drink a beer. A merely propositional imaginative project like this example has the advantage that it is relatively effortless and quick when details are relatively poor. There are no categorical reasons why propositionally imagining another person’s situation cannot be rich in detail. Merely propositionally imagining details of the overall mental state of a person is, however, not that easy because one has to think certain things through. For example, a simulator can propositionally simulate that she is angry. Whether this state comes with a certain motivational force or is accompanied by certain beliefs or desires would represent additional facts about the mental state. It requires a certain effort to merely propositionally simulate such additional facts.

Simulating being in another person’s situation can also be performed in an experiential way. The simulator might simulate what she would perceive if she were in the target’s situation, i.e. she might simulate what she would see, hear, smell and so on. The simulator may also simulate other states with experiential qualities she would encounter if she were in the situation, such as emotional states. By means of simulating experientially being in the target’s situation, the simulator can learn what it would be like for herself to be in the target’s situation. On the one hand, it seems that we need to be more focused and expend more imaginative effort to imagine not only a proposition but also the experiential qualities of certain states. On the other hand, however, if a simulator succeeds in simulating certain mental states in this way, other aspects of the situation are much easier to simulate as a result. Think of the following example. One simulator simulates merely propositionally that she sees an aggressive-looking dog. Another simulator simulates experientially seeing an aggressive-looking dog. The content of this simulation is the mental image of a dog. A mental image contains lots of details about a dog, such as its size and the way in which it looks aggressive. Because of these details, we can say that this simulation is more vivid than a merely propositional one. If one simulates, in this way, the sight of an aggressive-looking dog, for many simulators it will require less effort to further simulate being scared in this situation. That certainly does not mean that merely propositionally simulating seeing an aggressive-looking dog would not allow for merely propositionally simulating being afraid of the dog. It just seems that, because of its richness and vividness, it is often easier to react to the mental imagery than to the mere proposition.

So far, we can distinguish between propositionally simulating and experientially simulating being in the target’s position. This distinction is of course not exclusive. The content of experiential simulations can and often has propositional elements. In the following, I use the term ‘merely propositional simulation’ to indicate simulation projects with only propositional content.

If the epistemic aim of the simulation is to learn which state the target is in, both forms of simulation that I have presented face a problem: people often react very differently to a situation. Where one person feels afraid, another might become angry. Exactly what one thinks, feels, or desires in a specific situation depends a great deal on their character traits, on how they have been shaped by past experiences, or even on their physical condition at this moment. For this reason, many philosophers have suggested that we have to engage in a rather different simulation project if we want to learn about another person’s mental state. Instead of putting ourselves into the other’s situation, we have to take on their perspective. In Robert Gordon’s words, an ‘egocentric shift’ in the imagination project is required (Gordon 1995). The imaginer transforms herself imaginatively into the target person and in this way imagines being the other in her situation. This means that we try to simulate the other person’s reactions by considering their known character traits and so on.

We have distinguished between merely propositionally and experientially imagining being in the target’s situation. This distinction can also be applied to imagining being the target in her situation. This leaves us with a matrix with four fields:

Table

	Imagining being in the target’s situation	Imagining being the target in her situation
Merely propositionally imagining	Merely propositionally imagining being in the target’s situation	Merely propositionally imagining being the target in her situation
Experiential imaging	Experientially imagining being in the target’s situation	Experientially imagining being the target in her situation

We are, however, confronted with a problem here, and perhaps even with a dilemma: because of the complexity of human psychology, it seems that simulating our own reaction in a situation is not a reliable way of learning about the target person’s reactions. For this reason, an egocentric shift seems necessary if we want to predict another person’s psychological reactions or behavior. Imagining being another person is problematic, however, in several ways. First, it seems that the imaginer needs knowledge about the target person, about her personality, character traits, and so on. Although this knowledge is not at the level of a psychological theory, the process of imagining being another person cannot be a knowledge-poor process, as simulation theorists originally claimed. This problem could be solved by simply accepting that certain simulation projects are not as knowledge-poor as claimed (see Spaulding 2016; 2017; 2018). Furthermore, many philosophers are skeptical about the possibility of imagining being another person at all (see Goldie 2011; for discussion, see Langkau 2021). One of Goldie’s arguments is that psychological reactions are somehow often based on unconscious mental states and mechanisms which are not transparent to us. Unconscious states or mechanisms, however, cannot be part of a simulation project or, if they are, then they are not unconscious or subconscious to the imaginer (see Goldie 2011). Two further obvious problems are, first, that we will in practice often lack sufficient knowledge of another person. Second, it is difficult to quarantine the simulation process from other mental states and processes in such a way that they do not influence the simulation at all. This becomes problematic, especially in cases where the imaginer’s and the target’s personalities differ significantly. It will be more difficult to imagine psychological reactions which are very different from the imaginer’s own reactions. For these and perhaps even more reasons, it is quite likely that the imaginer will fail in the project of imagining being the other person.

One way of solving the problem is to see the two types of imaginative projects – (a) simulating being in the other person’s situation and (b) being the other person in her situation – as two ends of a continuum, with stages in between. An imaginer might fail to imagine being the other person. Nevertheless, she can try to feed her imaginative project with as much relevant information about the target as is available to her (for a developed account of empathy as perspective-taking beyond the distinction between (a) and (b), see Maibom 2022). The closer the imagined person is to the real target, the better the chances that the simulated reaction will be the same as or similar to the target’s real reaction. When I use the term ‘simulating being the other person in her situation’, I understand this in the weak sense as a simulation project which sufficiently considers information about the target person. In contrast, the term ‘simulating being in the other person’s situation’ will refer to simulation projects which consider no further information about the target person, her personality, and so on.

4. Simulation as reason to disbelieve the speaker

In the literature on social cognition, simulation processes are often seen as means of predicting the target person’s psychological reaction or behavior. However, people do not only use simulation in situations where there is no other source of knowledge about the target’s state. Often, the target can tell us what she thinks or feels but we simulate her state nevertheless. Here is one example:

(sexist comment) A woman reports that she felt uncomfortable, afraid even, in a meeting with only male colleagues when some of them started making comments about the way she was dressed.

This example presents a case of special testimonies, namely testimonies about the speaker’s own experience and her emotional reactions in particular. We can easily think of other people who try to understand the person in the example or who at least form opinions on what is reported. One possibility is that hearers do not believe what the person reports about her experience. In the case of the sexist comment, the hearer would not believe that the woman felt uncomfortable and even afraid because of her colleagues’ comments. One explanation for the hearer’s disbelief could be along the lines of Fricker’s analysis of testimonial injustice: identity prejudices against women could be the reason why the hearer in this case does not believe what was reported and could silence the speaker in different ways.

I follow here many authors who have argued that not being believed or heard or being silenced in such a case constitutes a serious harm for the speaker (for example Collins 2000; Dotson 2011; Fricker 2007; Langton 1993; Tuana 2006). If the hearer does not believe that the speaker reacts in the way she did he could assume that she mistakenly reported about her emotional experience. It is possible that a person makes mistakes in identifying her emotions. However, it is not very likely that a person believes that she was afraid also she was amused or felt flattered by the comment of her colleague. Believing that the speaker is not able to identify her own emotional experiences correctly means to ignore her expertise of her own experiences and to ignore her cognitive capacity to reflect on her own experiences. The hearer’s doubts concern thus very existential epistemic capacities. I will assume that the harm that results from credibility deficits in this case can be analyzed in similar ways to those Fricker describes (Fricker 2007, 44ff.)

The speaker’s emotional reaction toward the sexist comment is also one reason to identify it as such. If a hearer does not believe that the speaker was afraid or felt uncomfortable it is also likely that he will not accept that the sexist comment was sexist or in general a morally problematic act. Not being heard or believed means then in the sexist comment case that oppressive behavior will not be recognized as such by the hearer(s). The speaker will as a result not be able to successfully accuse her colleagues (see Langton 1993).

In the case I will analyze, identity prejudices against the speaker will not prevent the hearer from listening. This hearer instead tries to understand the speaker by simulating. A hearer in the sexist comment case can try to simulate the speaker’s experience in different ways.

First, the hearer can engage in an experiential simulation project. Relevant for the sexist comment case is that, in general, it is difficult to experientially simulate emotional reactions which differ from the simulator’s own reactions, i.e. the emotional reaction the simulator would have in the target’s situation or the emotional reactions the simulator believes she would have. If the hearer would not feel uncomfortable or even afraid if he were in the speaker’s situation, it is difficult to experientially simulate both being in the target’s situation and having these emotional reactions and being the target in her situation who has the reported emotional reaction. It is then very likely that attempts to experientially simulate being in the target’s situation or being the target in her situation fail, which means that the simulation gets interrupted or cannot even be started.

Second, a hearer can engage in a merely propositional simulation project. Merely propositionally simulating being in the speaker’s situation takes less effort than experiential simulation projects. Because it is relatively effortless, we can assume, first, that a lot of people engage in this sort of simulation project and, second, that this project will not fail as such. Because it is the most effortless simulation project, it is very likely that a simulator will only merely propositionally simulate being in the target’s situation instead of simulating being the target in her situation. The more simulator and speaker differ with respect to their personality, age or character traits and social background, the more likely the simulated emotional reaction will differ from the reaction reported by the speaker.

Now, let’s imagine that the simulator is a man who has no experience of being a victim of sexist behavior. We can assume that the emotional reactions in the case of the sexist comment do not only depend on personality and character traits. The social and experiential background will in this case be of great importance. A man is less likely to be reduced to a sex object or to become a victim of sexual violence. It is less likely that a man will have been ignored in his professional context or will have faced any other disadvantages simply because of his gender. If a man simulates being in the speaker’s situation, he simulates being a man in the woman’s situation. In the sexist comment example, this means exchanging the woman’s specific perspective with a man’s perspective in the simulation. With a man’s background and perspective, he is more likely to interpret the comments as compliments, for instance, and not to interpret them as examples of aggressive or otherwise problematic behavior. Seeing the comments from a man’s point of view, as compliments without any hint of threat or disrespect, makes it very likely that the simulated emotional reaction will be positive.

In general, for merely propositional simulation of being in the target’s situation, it is very likely that if the target’s situation is an instantiation of oppression and the target is a member of the oppressed group and the simulator is not, the simulated psychological reaction will differ from the target’s real psychological reaction in her situation. In these cases, the differences between the simulated reactions and the reported reactions are not incidental, but persistent and systematic because of the speaker’s and hearer’s different social background and it’s influence on their emotional reactions towards the act of oppression.

However, it is important to acknowledge that, for a man in the sexist comment case, simulating being in the speaker’s situation is not morally problematic per se. On the contrary, the simulation could help to identify and to acknowledge the different emotional reactions.

It is also very important for my argument that we can assume that this person either has no relevant identity prejudices against women or, if they do, that these prejudices do not play a causal role here. The literature on empathy and simulation suggests that simulation is not only a means of predicting the future behavior of other people, but also a way of trying to understand them. It is suggested that this happens regularly and it is not assumed that hearers only engage in simulation projects when they already have a reason not to trust the other person (Stueber 2006; Goldman 2006). For this reason, I will assume that our hearer in the sexist comment case simulates as a regular attempt to understand a speaker. In particular, we don’t need to assume that identity prejudices are the reason to engage in the simulation project in this case.

Hence, we have two scenarios: in the first, a hearer tries to experientially simulate but fails. The simulation project is interrupted or it does not take place at all. In the second case, the hearer is successful in performing the merely propositional simulation project as such. The simulator, however, simulates being himself in the target’s situation and simulates an emotional reaction which differs from the reported reaction. At best, the simulator can learn how he would react in the target’s situation, but he cannot learn how the speaker reacts.

5. Reckless or negligent ways of dealing with the simulation

The hearer has several ways of dealing with the simulation, and some are epistemically and morally problematic.

First, Nomy Arpaly argues that failed imagination projects can result in a credibility deficit on the speaker’s side: because the simulator cannot imagine having the emotional reaction that is reported by the speaker, the simulator does not believe what the speaker reports about her experience (Arpaly 2020). As we have seen, the chances that experiential simulation projects fail are high, particularly if the reported emotional reaction differs from how the simulator would react or thinks he would react. So, in the case where the simulator tries to experientially simulate, the simulator does not trust the speaker, presumably because he thinks that the speaker’s reported reaction is not possible because he could not imagine having the reaction.

Second, the simulator could confuse the nature of the simulation project. He could believe that he simulated being the other person even though he simulated being himself in the other person’s situation. On the basis of this confusion, the simulator then forms the belief that what he simulated is how the target person really reacts and thus believes that the target did not report her experience truthfully.

In the following, however, I will not focus on these two cases but on a third possible problem, arguing that we must interpret this case as either reckless or negligent and hence as a moral wrong.

In the case I want to focus on, the simulator who successfully propositionally simulated being in the speaker’s situation is aware of the fact that he was simulating being in the target’s situation (and not being the target in her situation) and hence is aware of the fact that the simulated reaction would at best be his reaction in the target’s situation. It is fair to assume that many people are successful in such a simulation project and that they are not confused about what they have simulated. Hence, we can assume that by excluding the first mentioned cases we are still in no danger of talking about an insignificant number of cases.

In the example, the simulator could ignore the fact that people react differently. By ignoring this fact, the simulator could believe that the simulated reaction is the only possible reaction. Because the target person reported a different reaction, the simulator could then, for example, form the belief that the target person could not have told the truth about her emotional reaction in the situation.

Obviously, the premise (that the simulated reaction is the only possible reaction) in this reconstructed argument is false because people do not react in the same way. Even if we assume that it is true that the simulator would react differently if he were in the speaker’s situation, he has no reason to believe that another person cannot react in a different way. Hence, the simulation does not provide any reason to distrust the speaker.

Second, in a real situation, people often do not react quite as they predicted on the basis of their simulation. Heidi Maibom presents several psychological studies which indicate that people tend to simulate how they think they should react, but not how they really react (Maibom 2016; 2018). The mechanisms that lead us to simulate in a way that is different from how we would indeed react in an actual situation are manifold. One reason is the nature of the merely propositional simulation. These simulations are often exercised quickly and with little effort. For this reason, merely propositional simulations are often, though not necessarily, poor in detail. Details of the person’s situation can, however, have an influence on how she (emotionally) reacts in a situation. If we do not simulate a situation with sufficient details, it can easily happen that we simulate a reaction different from that which we have in reality. If we assume that, in our case, the simulation is very poor in detail, then the simulation is therefore not trustworthy. The chances that the simulated reactions differ from those the simulator would really have would simply be too high. For this reason, from a very poor simulation project, the simulator cannot even learn anything about his own emotional reactions in a specific situation.

The conclusion that the speaker did not tell the truth does not follow, therefore, and we can easily see the epistemic problem of the simulator’s thoughts here. However, there is also something morally wrong in these cases on account of the fact that the simulator is reckless or negligent.

First, if the simulation was performed quickly and effortlessly, the simulator trusts a quick and effortless simulation project more than another person’s testimony about her experience. He ignores the fact that his simulation was just hasty and weak. The simulator does not consider that he did not try to pay more attention and put more effort into the simulation project. He should be aware of the poor nature of his simulation project and thus realize that his simulation is not trustworthy and thus provides no reason – or at least no good reason – not to trust the speaker. Nevertheless, there is an additional problem.

The further problem is that the simulator does not consider the obvious fact that people can react differently. We can assume that an adult person knows this fact from experience: while one person might enjoy watching horror films and finds them amusing, another does not enjoy them at all and instead finds them frightening. One person might be disgusted by a spider, while another is fascinated. Everybody has experienced such differences and so we can assume that adults possess knowledge about these differences or that it is at least easily accessible for them.

The sexist comment example points to another important aspect: although the speaker’s personality here has or can have an influence on how she reacts emotionally, it is important to acknowledge that this speaker is a member of a group that regularly faces sexist behavior. The speaker’s specific perspective as a member of an oppressed or marginalized group is important in the examples (Collins 2000; Wiltsher 2021). There will be at least some cases in which a hearer knows or can know about the influence of social and experiential background or specific perspectives on the way we react emotionally.

By considering that people can react differently and, moreover, by considering that the female speaker’s experiential background plays a role in her emotional reaction, the simulator could see that his belief about his own possible reaction based on his simulation is no reason not to trust the speaker’s report. Because he fails to consider this knowledge he falsely takes his simulation as a reason not to trust the speaker.

Not considering knowledge which is relevant in the situation can be a reckless decision. It can also be a case of negligence. Santiago Amaya highlights the epistemic dimension which is at the core of negligence in general:

The negligent person has enough information to act in line with certain moral standards or, at least, has easy access to information that would allow her to do so. But because she fails to advert to that information or to the need to seek it, she winds up acting below those standards. What she does, therefore, is not a reflection of what she intends or how willing she is to do or to risk doing things with potentially bad consequences. It is a rather a reflection of her failure to see her actions (or her omissions) in the light of some available information at the time. (Amaya 2022, 8)

This is exactly what happens in terms of the negligence in the sexist comment example: the simulator negligently fails or omits to consider his relevant knowledge and the impact on his stance toward the speaker.

Identity prejudices may of course be the reason for the hearer’s/simulator’s recklessness or negligence, but they need not be the reason. Consider the following example where we have no reason to assume that identity prejudices play a role:

(Sunday morning) Speaker and hearer are cousins, and both male. The speaker reports that he enjoyed running on Sunday morning. The hearer simulates being in this situation and simulates that he would not enjoy this run at all.

As in the sexist comment case, the simulator/hearer ignores the obvious fact that people react differently and, based on his simulation, forms the belief that the speaker does not report his experience truthfully. In the Sunday morning case, we have no reason to assume that the hearer has any identity prejudices against the speaker. Hence, we cannot assume that identity prejudices are the reason why the hearer behaves recklessly or negligently in this case.

We can then assume that sometimes people do not consider the fact that others might react differently, independently of the social background of the speaker or identity prejudices. If it is true that this recklessness or negligence can be explained in some cases without reference to identity prejudices, it seems possible that identity prejudices do not have to be the reason in examples like the sexist comment case either.

7. Is the hearer’s/simulator’s recklessness or negligence an exclusively epistemic error?

At this point, one might still not be convinced that the simulator in our example – who has no relevant (or causally relevant) identity prejudices – performs more than an epistemic mistake. Hence, one might still think there is nothing morally wrong in this case.

Fricker discusses a case – which looks prima facie very similar to our sexist comment example – where she claims that no moral wrong is done. Let’s take a look at her example. A philosopher is involved in a discussion about a philosophical topic, namely moral fictionalism. However, the philosopher believes that her interlocutor is a medic rather than a philosopher, and does not give him the credibility he deserves. In this case, the philosopher has a false belief as a result of some hasty and careless internet research about the other person. Fricker characterizes this as an ‘ethically innocent but epistemically capable error’ (Fricker 2007, 22) and claims that the case does not constitute a case of testimonial injustice.

One might be inclined to think that a false belief based on careless internet research is relevantly similar to what went wrong in the sexist comment case and the propositional simulation. It is, however, very important that Fricker describes her scenario in a way that makes it clear that no real damage is done to the interlocutor. The conversation was probably only a casual chat, and certainly nothing important like a job interview. And we can understand the example to imply that the philosopher attributing less credibility to the interlocutor’s words than if she had believed him to be a philosopher has no further negative consequences for him. The credibility deficit in Fricker’s example cannot be linked to other forms of structural injustice either, as would be the case in a proper example of testimonial injustice. In particular, Fricker describes the case in such a way as to make us think that the philosopher will change her mind and perhaps even feel embarrassed as soon as she learns of her mistake (Fricker 2007, 22).

This is entirely different in the sexist comment example. In this case, the speaker suffers from a credibility deficit which is highly problematic for her and she clearly suffers from negative consequences. Because of the harm done to the speaker, the hearer’s/simulator’s recklessness or negligence is neither a purely epistemic nor an innocent mistake.

Further, we can now also see why it is very likely that a person from a marginalized or oppressed group will suffer from an unjust credibility deficit when she reports her emotional experiences if the hearer is not from (the same) marginalized or oppressed group and is reckless or negligent in the way I have described. Although the injustice does not occur because of identity prejudices on the hearer’s side, the injustice is also not merely incidental, but rather persistent and systematic.

We can now see why structures of oppression influence such cases and from where the moral wrong stems. The experiential background does not shape both the speaker’s and the hearer’s reactions to the (unrecognized) oppressive behavior coincidentally, but because they are members of different social groups and in at least one relevant respect belong to an oppressed and a non-oppressed group, respectively. Hence, the emotional reactions towards an act of oppression simulated by the hearer will differ systematically and persistently from the speaker’s reactions.

This is in itself not problematic or morally wrong. It becomes problematic when the hearer deals with the simulation in a reckless or negligent way. This recklessness or negligence does not need to be motivated by identity prejudices against the speaker and other members of her group, but as recklessness or negligence, it is still morally wrong.

We can see the relevant differences between the sexist comment and the Sunday morning case. Let’s assume for the sake of the argument that the credibility deficit in the latter case constitutes, somehow, a serious harm for the speaker. We might even be able to tell a story where the hearer persistently simulates different emotional reactions than the ones reported. However, the harm in the Sunday morning case cannot be connected to other forms of structural injustice. Hence, the case lacks the feature of being a systematic injustice in Fricker’s sense.

In Fricker’s central cases of testimonial injustice, we can track both the moral wrongness and the persistent and systematic features to the hearer’s identity prejudices against the speaker. In the case I have analyzed, we have to track these features along two different paths. The moral wrongness stems from the hearer’s recklessness or negligence, but the injustice done to the speaker inherits the systematic features and its persistent nature from the simulation. Thus, the unjust credibility deficit can be connected systematically to other forms of social injustice suffered by the speaker or other members of her social group. Hence, in such cases, reckless or negligence ways of dealing with simulations will systematically and persistently lead to unjust credibility deficits.

8. Conclusion

The aim of this paper has been to focus on a case of testimony about emotional experiences which is in many respects relevantly similar to Fricker’s central case of testimonial injustice, but differs in one important respect: I have assumed that the hearer has no identity prejudices against the speaker or that such prejudices are not causally relevant in the case. If my analysis is correct, we have identified additional cases of unjust credibility deficits to those discussed in recent literature. It goes without saying that I do not want to argue that identity prejudices can never play a causal role in simulation projects or the way simulators deal with them.

In the case on which I focused, the hearer simulates being himself in the speaker’s situation. Because he is not a member of the speaker’s social group, a marginalized or oppressed group, the simulated emotional reactions will differ from the reported emotional reactions. I have argued that this result is systematic in the sense that it is related to the hearer’s and the speaker’s respective social backgrounds. Because of the different experiential backgrounds, we can assume that the difference between the simulated and the reported emotional reaction is not incidental but persistent.

The simulation is in itself not morally problematic. On the contrary, if a hearer acknowledges the differences, it could be an epistemic advantage and a step toward understanding the speaker better. However, I have focused on a case where the hearer behaves in a rather different way. He either recklessly or negligently fails to consider knowledge about the differences in people’s emotional reactions and comes to believe that his simulated reaction is the only possible reaction, thus believing that the speaker does not report her emotions truthfully. I have argued that, in this case, the ‘epistemic poison’ results from the recklessness or negligence. This recklessness or negligence is not necessarily due to an identity prejudice against the speaker. This credibility deficit is unjust, yet its persistent and systematic nature stems from the fact that a hearer who is not from the same marginalized or oppressed group as the speaker will simulate different emotional reactions to those reported by the speaker. Hence, the unjust credibility deficit can be connected to other forms of social injustices suffered by the speaker.

Acknowledgements

For helpful comments I want to thank Katharina Anna Sodoma, Lizzy Ventham, Thomas Schramme and Neil Roughley, the participants of the conference ‘Interpersonal Understanding and Affective Empathy’ in Liverpool, held 26–28 June 2023, in particular Amy Kind, Moritz Müller and the participants of his research colloquium at the University of Tübingen and two anonymous referees. The conference and the work on this article were generously founded by DFG and AHRC.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Christiana Werner

Christiana Werner is a Postdoctoral Research Assistant at the University of Duisburg-Essen, Germany. Her research focuses on empathy, imagination and understanding in the context of philosophy of mind, esthetics and epistemology.