2. Rape myths

Rape myths can be understood as “attitudes and generally false beliefs about rape that are widely and persistently held, and that serve to deny, justify, and tolerate male sexual aggression against women” (Lonsway and Fitzgerald Reference Lonsway and Fitzgerald1994: 134). In this section, I defend the view that rape myths function to deny, justify, and tolerate sexual violence against women. However, I take issue with versions of this view that ignore the deadly seriousness with which some accusations are met.

Consider the following account of rape myths offered by Hall et al. (Reference Hall, Howard and Boezio1986). They attempt to capture the functioning of rape myths by providing three categories that rape myths can be sorted into: (i) myths that deny the existence of rape or the scope of the problem, (ii) myths that excuse the behavior of the rapist, and (iii) myths that deny that rape is a serious crime. Here are some common rape myths, sorted into the three categories provided by Hall et al. (Reference Hall, Howard and Boezio1986):

1. i. Myths that Deny the Existence of Rape

   1. a. Women routinely lie about rape.Footnote ²

   2. b. Women claim they've been raped when they regret the sexual interaction.

   3. c. Women could resist rape if they really wanted to.

   4. d. Consent cannot be withdrawn after it has been given.

   5. e. Consent is automatically present if a sexual act has occurred between the same parties before.

   6. f. “Attractive” men don't rape women, and “unattractive” women are never raped.

2. ii. Myths that Justify the Rapist's Behavior

   1. a. Women who were raped were “asking for it”, either by dressing promiscuously or drinking excessively.

   2. b. Rape is a crime of passion.

   3. c. If he was drunk, he couldn't help it.

3. iii. Myths that Deny the Seriousness of Rape

   1. a. Only promiscuous women are raped.

   2. b. Most women secretly want to be raped.

   3. c. Most rapists only rape once.Footnote ³

This framing of rape myths emphasizes minimizing myths. It suggests that rape myths function to deny, justify, and tolerate sexual violence against women by perpetuating the idea that rape is “not that bad”. On this account, the myths that cast rape as a rare occurrence both obscure the danger women face and trivialize their fear; the myths that deny the seriousness of rape encourage minimal, if any, punishment; and the myths that justify the rapist's behavior shift the blame to the victim and so minimize the violation. This way of fleshing out how rape myths deny, justify, and tolerate sexual violence against women remains popular (see, for example, Roxane Gay's ‘Not That Bad’ (Reference Gay2018)). However, this analysis is seriously flawed.

First, it seems to ignore the contemporary (and historical) seriousness with which allegations that black men have raped white women have been met. Not only are these reports often deemed credible, they are deemed credible reports of serious wrongdoing. To return to the example from the Introduction, when it was (falsely) claimed that black men were raping white women at horrifying rates after the Reconstruction Era, it was said that this crime “caused civilized men to revert to the original savage type whose impulses under such circumstances have always been to ‘kill, kill, kill’” (Davis Reference Davis1981: 187). This was offered up as justification for the brutal lynchings – white men were, supposedly, driven to the extreme violence by their moral outrage.Footnote ⁴ The mythical black rapist is not – and was not – real. But the success of the narrative suggests that rapes were not taken to be “not that bad”; on the contrary, the success of the narrative seems to have depended on rape being considered a seriously morally reprehensible crime.

More generally, it is hard to square the view that rape culture persists because people think rape is “not that bad” with the widespread disavowal of rape. Rape features prominently in TV crime shows, and viewers are expected to root for the perpetrator to be brought to justice; graphic rapes of young college women by student-athletes are featured on the news, to nationwide outrage; while frat boys at Yale may hold up signs reading, “no means yes and yes means anal” (Orbey Reference Orbey2018), they are met with broad condemnation. In general, people believe that rape is bad.

However, while there may be widespread disavowal of rape, it is still apparent that sexual violence against women is largely carried out with impunity. One in six women is the victim of rape or attempted rape (and the statistics are worse for trans women and women of color, especially Indigenous women) (“Victims of Sexual Violence: Statistics”, RAINN); only five out of every 1,000 rapes result in a conviction (“The Criminal Justice System: Statistics”, RAINN Reference RAINN2017); Donald Trump was elected despite his claims to “grab [women] by the pussy”; Brett Kavanaugh was confirmed to the Supreme Court in the face of credible allegations that he sexually assaulted Christine Blasey Ford while they were in high school. It is hard to ask for clearer evidence that sexual violence against women is widely tolerated. So, there is clearly a sense in which people are failing to take sexual violence seriously. How do we make sense of this failure, given the public disavowal of rape?

To solve this puzzle, we need to attend to the category of rape myths that is overlooked by Hall et al. (Reference Hall, Howard and Boezio1986): catastrophizing myths.

1. iv. Myths that Catastrophize Rape

   1. a. Rape always involves overwhelming physical force.

   2. b. Rape occurs in dark alleys late at night.

   3. c. Rape is committed by strangers, not by people close to you.

   4. d. Rape is committed by people incapable of respecting or caring for women.

These myths, unlike the previously discussed myths, condemn rape as a monstrous act – but they still are not based in reality. Women are significantly more likely to be raped by someone close to them than they are to be raped by a stranger; rape does not always involve any – let alone overwhelming – physical force; and many people who commit rape are perfectly capable of forming respectful, even loving, relationships with other women (Yap Reference Yap2017: 4–5).

There seems to be tension between minimizing rape myths and catastrophizing rape myths. One possible interpretation of the seemingly contradictory set of rape myths is that people simply hold contradictory beliefs about rape. They believe simultaneously that rape always involves overwhelming physical force, and that women could fight off rape if they really wanted to; they believe simultaneously that rape is a vicious attack in a dark alley late at night, and that rape is really not that bad. On this interpretation, some rape myths work to trivialize, justify, or deny sexual violence against women, while others do not – people simply have mixed-up and incoherent beliefs about rape. Thus, people might sincerely condemn rape while still rooting for Trump or Kavanaugh – their beliefs simply don't cohere.

Alternatively, we might suggest that there's no incoherence in people's beliefs at all – people simply condemn some rapes, while tolerating others. People might, for instance, think that rape is morally intolerable when the victim is a white woman and the perpetrator is a black man, but think rape is tolerable if the races are reversed – that is, if the victim is a black woman, and the perpetrator is a white man. So, rather than positing that people have incoherent beliefs, this explanation suggests that people have grossly immoral beliefs.

Neither of these routes is the one that I will take. Both of these explanations suggest that, in some way, tolerant attitudes toward rape are somehow playing an important role in sustaining rape culture. I maintain that rape myths primarily function to deny, justify, and tolerate sexual violence against women, but I deny that this is done by cultivating tolerant attitudes towards rape.Footnote ⁵

I suggest that what is endemic to contemporary rape culture is a systematic failure to recognize many instances of rape. Too often, people incorrectly dismiss a situation as not really rape. This dismissal doesn't depend on having a tolerant attitude towards rape. Indeed, as I hope to show, people's catastrophized conception of rape can be the driving force behind their unjust dismissal of rape allegations.

Rape myths enable people to dismiss various rape allegations by denying that there was really rape in two ways. First, catastrophizing rape myths distort our assessment of what “real” rape looks like. “Real” rapes are characterized by the catastrophizing rape myths – “real” rapes are violent and terrifying; they happen to innocent women who are alone at night, and who could not possibly fight off their attacker. “Real” rapes, like those portrayed in TV crime shows or on the news, are obviously heinous and must be condemned. In contrast, the myths that minimize rape do so by characterizing such violence as not really rape. If she was wearing a short skirt, she must have wanted it – not real rape. If it didn't require overwhelming physical force, she could have fought him off – not real rape. If she has said yes to him before, she probably said yes to him this time – not real rape. Since these instances don't fit the stereotype of “real” rape, one can dismiss them while remaining consistent with sincerely held “anti-rape” beliefs and values.

A catastrophized conception of rape sets demanding standards. For something to count as “real” rape, we need to be able to conceptualize what happened as being a moral catastrophe, to see the victim as vulnerable to serious violation, and to see the accused as capable of doing something monstrous. Audrey Yap makes a version of this point when she argues that a narrative that casts rapists as necessarily monstrous gives rise to a problematic modus tollens inference that exonerates many people we see as non-monstrous:

1. a) If someone commits a sexual assault, then that person is a monster.

2. b) Person X is not a monster.

3. c) Therefore, Person X did not commit sexual assault (Yap Reference Yap2017: 16).Footnote ⁶

We can generalize this point: a catastrophized conception of rape gives rise to an exonerating modus tollens inference any time we fail to see the violation at hand as a catastrophic violation. This could happen for all kinds of reasons. It might be because we are incapable of seeing the accused as doing something monstrous, but it might also be because we're incapable of empathizing with the victim and recognizing the severity of the violation she has endured, or because we think that what she's reporting just doesn't seem that bad.

The upshot of this is that, somewhat surprisingly, a catastrophized conception of rape can result in the unjust dismissal of credible rape allegations, rather than a commitment to taking such allegations seriously. Thus, the catastrophized myths are an important part of contemporary rape culture.

Skewing our conception of “real” rape is just one of the ways rape myths distort the epistemic resources that we bring to bear on rape allegations. The second distortion involves positioning women who come forward as likely to be lying. This is the work of myths that, in Hall et al.'s (1986) terms, “deny” sexual violence against women – these myths deny such violence by suggesting that women often lie about rape. When the victim is judged to be untrustworthy, it might be agreed that what is reported to have happened rises to the level of “real” rape – but what really happened does not. What “really” happened might be that the woman was broken up with, or had sex that she now regrets, or even nothing at all. In these cases, allegations are written off on the grounds that the reported rape is fabricated, and so not “real”. This, again, allows people to insist that they condemn rape, while justifying and tolerating individual rapes – they fail to see those rapes as really rape.

So, on my view, contemporary rape culture tolerates sexual violence against women not by cultivating tolerant attitudes towards rape, but by distorting our conceptual resources in ways that allow us to (wrongly) dismiss most rapes as not really rape. This explanation of how rape myths distort our epistemic resources explains how there is widespread, public disavowal of rape and that the vast majority of rapes are carried out with impunity: people fail to recognize that the vast majority of rapes are rapes. Crucially, though, there is a small subset of rapes that people vigorously condemn. This explains the deadly seriousness with which some allegations are met. “Real” rape is considered to be a serious – even unthinkable – moral infraction. And while this moral infraction is so serious that many perpetrators will be deemed incapable of the infraction, not everyone who is accused is held in such high regard – there are some people who we do think are capable of monstrous moral infraction. This, of course, is how we end up with such a racialized response to rape allegations – it is (unjustly) easier for us to imagine a black man committing such a heinous crime.Footnote ⁷

The upshot of this is not that we need to stop thinking that rapes are bad. Of course rapes are bad. The mistake is, instead, with thinking that rape is necessarily monstrous or terrifying or somehow inhuman, and that it is because of these features that rape is bad. The reality is this: people we know and like commit rape; this rape is bad. Rape occurs without physical violence; this rape is bad. Rape occurs and the victim never leaves the perpetrator; this rape is bad. By catastrophizing rape, we make it more difficult to see most rapes for what they are, and to recognize what it is that makes these violations so harmful.

3. Rape myths and epistemic injustice

The dual distortions of the epistemic resources we bring to bear on rape allegations fit neatly into the classic types of epistemic injustice identified by Fricker: hermeneutical injustice and testimonial injustice. These types of epistemic injustice are not my focus; I aim to draw attention to another phenomenon at play here. However, it will help to get these types of injustice on the table. Doing so will help me to distinguish the kind of injustice I will identify, and also to highlight an important terminological difference between Fricker's account and my account.

Hermeneutical injustice concerns the role prejudice can have in determining the conceptual resources that are developed to make sense of our experiences. Fricker defines it as “the injustice of having some significant area of one's social experience obscured from collective understanding owing to hermeneutical marginalization” (Fricker Reference Fricker2007: 158). When hermeneutical injustice occurs, certain social experiences are incomprehensible because people lack the interpretive tools necessary for making sense of those experiences, and they lack these tools as a result of the exclusion of marginalized people from dominant meaning-making practices. The distorted, overly narrow conception of rape that rape myths impart is a clear instance of this kind of injustice – the vast majority of rapes are seriously misunderstood because people lack the interpretive tools to make sense of these rapes as rape (see Jenkins (Reference Jenkins2017), for a detailed discussion of this point).

Rape myths’ maligning of victims’ trustworthiness is a clear instance of testimonial injustice. Testimonial injustice occurs when “prejudice on the hearer's part causes him to give the speaker less credibility than he would otherwise have given” (Fricker Reference Fricker2007: 4). Not taking an account seriously because it was given by a woman; not taking an account seriously because it was given by someone with an accent; not taking an account seriously because it was given by a black person; these are all examples of testimonial injustice. Rape myths – which are plainly prejudicial – can result in the unjust dismissal of rape victims’ testimony. Thus, this seems like a relatively clear instance of testimonial injustice.

This brings me to the important terminological distinction. Fricker characterizes testimonial injustice as an unjust credibility deflation. However, I use the term ‘credibility’ slightly differently than Fricker does. I follow Karen Jones, who follows Locke, in understanding ‘credibility’ more generally.

Jones identifies two grounds of credibility: the trustworthiness of the testifier, and the plausibility of what is testified to in light of our background knowledge. The trustworthiness of the testifier is determined by considering two factors: “competence with respect to the truth of p and veracity with respect to [the] exchange” (Jones Reference Jones, Antony and Witt2002: 156). A testifier's competence with respect to the truth of p is determined by examining their “counterfactual tendencies”; competent testifiers will tend to believe some claim p only if p is true, and they will not tend to believe p if p is not true.

A testifier's veracity with respect to the exchange is determined by looking at her “local trustworthiness”. A testifier is locally trustworthy if, in exchanges relevantly similar to the one occurring, she would be honest (Jones Reference Jones, Antony and Witt2002: 156). If we have reason to doubt that she would be honest in relevantly similar exchanges (if, for example, the testifier would obviously benefit from lying), then she is not locally trustworthy and so not a trustworthy testifier. Trustworthy testifiers are, then, those people who are both locally trustworthy and competent with respect to their claim.

The second ground of credibility – plausibility in light of background knowledge – concerns how plausible a claim is. The more “astonishing” a case is, meaning the less it coheres with our background knowledge, the less plausible it is, and so the less credible it is (Jones Reference Jones, Antony and Witt2002: 157).

On this understanding of ‘credibility’, both testimonial injustice and hermeneutical injustice result in unjust credibility deflations. This is straightforward with testimonial injustice: testimonial injustice involves an unjust trustworthiness assessment, which results in a lower overall credibility assessment. It is slightly less straightforward with hermeneutical injustice – hermeneutical injustice is not primarily an unjust plausibility deflation. Instead, hermeneutical injustice has downstream effects on what we think is plausible: if we lack the conceptual resources necessary to make sense of a particular experience, reports of that experience will be more difficult to understand, and so will seem less plausible. Suppose, for example, that someone thinks a conventionally attractive man would never sexually assault a conventionally unattractive woman. When confronted with an allegation that alleges just this, we should anticipate that he will feel surprised; this is not how sexual assault works, according to the conceptual resources available to him. This will likely have a negative effect on his assessment of the plausibility of the allegation. Thus, hermeneutical injustice can also result in an unjustly deflated overall credibility assessment.

Given my more general use of ‘credibility’, I refrain from characterizing testimonial injustice as an unjust credibility deficit. Instead, I understand testimonial injustice as an unjustly deflated trustworthiness assessment. This is, I emphasize, merely a terminological divergence from Fricker's account – substantively, I understand testimonial injustice in the same way that Fricker does.

The two types of injustice identified by Fricker concern how prejudice can affect our assessments of the two grounds of credibility separately. What is important, for my purposes, is that our assessments of plausibility and trustworthiness can interact – we do not always make separate assessments of trustworthiness and plausibility.

Suppose, for instance, that I approach you and tell you that I am a sewer rat in disguise. The radical implausibility of this claim casts doubt on my trustworthiness with respect to it – I might be confused, pulling a strange prank, etc. In any case, it seems that the implausibility of my claim casts legitimate doubt on my trustworthiness. More realistically, consider the claim that anthropogenic climate change is a myth, a story conjured up by left-leaning scientists in order to motivate their desired policies. The implausibility of the climate denier's claim, again, seems to cast doubt on her trustworthiness with respect to it – she may be deliberately spreading misinformation, or simply enjoys provoking others, or has genuinely bought into this right-wing conspiracy theory, etc.

I take these two cases to illustrate that our assessments of trustworthiness and plausibility can interact, and that this interaction can be legitimate. The philosopher who has paid the most attention to this feature of our epistemic practices is Karen Jones. Jones recognizes that, if our assessments of the two grounds of credibility interact, unjust assessments of either ground can be amplified – an unjustly low assessment of trustworthiness or plausibility, if allowed to influence the assessment of the other ground, can result in further unjust credibility deflation.

I pursue this idea in what follows. While Jones raises an important issue, I disagree with her conclusion. Jones argues that we must make independent assessments of trustworthiness and plausibility, so that we can avoid further disadvantaging those who are unjustly deemed untrustworthy or whose reports are unjustly deemed implausible. I take a more moderate route: I maintain that assessments of plausibility and trustworthiness can legitimately interact, although this legitimate interaction has limits. I then explore how – and when – the violation of these limits amounts to an epistemic injustice, and use a case study to show how this epistemic injustice can help to flesh out our picture of the epistemic obstacles faced by victims who come forward about their rapes.

4. Jones and the Independence Rule

Jones argues that we must make independent assessments of the trustworthiness of a testifier and the plausibility of their claim (Jones Reference Jones, Antony and Witt2002: 157). Her argument centers around the following case study:

While Jones diagnoses several problems with Ferlise's reasoning, of particular importance for my purposes is her claim that one of the judge's failures was not keeping his assessments of the trustworthiness of the testifier and the plausibility of her claim separate.

Jones argues that an initial assessment that a testifier is untrustworthy can result in viewing her testimony through a “lens of distrust” (Jones Reference Jones, Antony and Witt2002: 159). This way of approaching testimony, Jones argues, can be self-fulfilling – it can lead us “to seek out evidence of inconsistencies, to magnify those we suppose ourselves to have found … to focus on them in our assessment of the story as a whole”, and to overlook possible explanations for discovered inconsistencies (Jones Reference Jones, Antony and Witt2002: 159). So, when we allow our trustworthiness assessment to influence our plausibility assessment, that influence can cause us to give a lower plausibility assessment than is actually warranted. Worse still, Jones maintains that allowing assessments of one ground of credibility to influence the other can result in “runaway” reductions to overall credibility – “our low initial trustworthiness rating leads to a reduction in the plausibility rating we would have given to the content of the story, and this in turn confirms our initial assessment of untrustworthiness, which in turn makes us only the more confident in our low plausibility ratings” (Jones Reference Jones, Antony and Witt2002: 160). When a testifier is subjected to runaway credibility deflations, there is not much she can do to get her interlocutor to believe her.

Jones argues that this dynamic can be seen in Ferlise's reasoning. Ferlise based his assessment that Kassindja was “incredible” on the following factors: she didn't know the current whereabouts of her mother; Ferlise doubted female genital mutilation was that much of a threat if she had been able to avoid it up to this point; and Ferlise found it surprising that she was able to survive in Germany with the help of two strangers (Jones Reference Jones, Antony and Witt2002: 160). As Jones notes, only the second consideration actually has any bearing on her asylum request. But more importantly, Jones thinks that these three factors could only lead someone to reject Kassindja's trustworthiness as a testifier, as Ferlise did, if these factors were viewed through a lens of distrust. Without the distortion caused by distrust, these three factors would not cause someone to deem Kassindja untrustworthy. Given Ferlise's history of lacking sympathy for female asylum seekers, it is plausible that he approached Kassindja's testimony with such an initial low trustworthiness assessment. This low trustworthiness assessment likely influenced his plausibility assessment, which then affirmed his initial trustworthiness assessment (Jones Reference Jones, Antony and Witt2002: 161).

So, Jones concludes that Ferlise's dismissal of Kassindja's testimony was the result of allowing his assessments of the two grounds of credibility to feed off of each other. This leads her to what she calls the Independence Rule: make independent assessments of the trustworthiness of the testifier, and the plausibility of their claim (Jones Reference Jones, Antony and Witt2002: 159). Specifically, what this rule forbids is: (i) allowing the implausibility of the present claim to influence the assessment of the testifier's trustworthiness with respect to it, and (ii) allowing the assessment that the testifier is locally untrustworthy to influence the assessment of the claim's plausibility. This means that a history of making implausible claims may be used as evidence that the testifier is not competent or locally trustworthy, although Jones does advise caution in making this inference (Jones Reference Jones, Antony and Witt2002: 158).

Jones specifies that a history of making implausible claims can still be used as evidence that a testifier is untrustworthy because she recognizes that, often, it seems like the plausibility of a claim does have some bearing on the testifier's trustworthiness. Even with this restriction, though, Jones’ claim seems too strong. Recall my hypothetical claim that I am a sewer rat in disguise, or the claim that climate change is a myth – the implausibility of these claims, I suggested, legitimately casts doubt on the trustworthiness of the testifiers. Jones’ Independence Rule would bar us from making these inferences, and so seems too strong.

That said, I do think Jones has picked up on an important flaw in Ferlise's reasoning – Ferlise unjustly allows his plausibility and trustworthiness assessments to feed off of one another, and this leaves Kassindja with a nearly insurmountable credibility deficit. This suggests that we need something to explain the difference between Ferlise's reasoning in Kassindja's case, and our reasoning about the climate change denier – why is one inference legitimate, but the other not?

My aim in the following section is to capture what's going wrong with the runaway credibility deflations that Jones is concerned with, without positing a rule as strong as the Independence Rule. I begin by briefly canvassing possible restraints on legitimate interaction between assessments of trustworthiness and assessments of plausibility. I then focus on the constraint I call the degree constraint and explain how the violation of this constraint captures what is going wrong with Ferlise's reasoning.

5. Legitimacy constraints

Let's focus on the climate denier. This example is meant to show that trustworthiness and plausibility assessments may legitimately interact – by saying something so implausible, the climate change denier casts doubt on her own trustworthiness with respect to her claim. Moreover, the implausibility of her claim seems to legitimately deflate our assessment of her trustworthiness with respect to other related claims. We might, for instance, also think she is not trustworthy when she praises the coal industry.

But there are limits to how far we can extend this deflated trustworthiness assessment. For example, the implausibility of her denial of anthropogenic climate change does not warrant a deflated trustworthiness assessment when it comes to claims about her children's whereabouts, or what she had for lunch – we do not have reason to think she is incompetent with respect to, or likely to be lying about, claims like this. I call this the Scope Constraint. The Scope Constraint reflects the domain-specificity of trustworthiness: I may be trustworthy in some areas, and not in others. Rarely are people universally untrustworthy.

In addition to the Scope Constraint, we might posit something like a Reasonableness Constraint: only initially reasonable assessments of trustworthiness or plausibility may legitimately influence your assessment of the other ground. The exact contours of this constraint will depend on your preferred account of reasonableness. I wish to remain neutral on these details; I suggest only that some version of a Reasonableness Constraint is plausible.

Finally, I want to posit a Degree Constraint. Even a reasonable plausibility or trustworthiness assessment only warrants some degree of influence on your assessment of the other ground. Some independent investigation into the trustworthiness and plausibility of the claim must be done. When the assessment of one ground of credibility determines the other ground, or otherwise exerts too much influence, the Degree Constraint is flouted. The violation of this constraint can capture what is going wrong in the runaway credibility deficits that Jones is concerned with – your initial low trustworthiness assessment causes you to give a lower plausibility assessment than you otherwise would have, which confirms your initial low trustworthiness assessment, and so on. By using the low trustworthiness assessment as (conclusive) evidence that the claim is implausible, no independent investigation into the plausibility of the claim is carried out. The same goes when the implausibility of a claim is used to conclude that a testifier is untrustworthy; the implausibility of the claim is taken to settle the matter all on its own. This is epistemic laziness that is not permitted by the Degree Constraint.

This list of constraints isn't meant to be exhaustive, but to give a sense for the limits on legitimate interaction between assessments of trustworthiness and plausibility; there are, presumably, further constraints that could be provided, and the details could be more fully worked out. This rough sketch is enough for my purposes; what matters is that when we flout restraints like this, we make an epistemic mistake. This mistake consists in making it too difficult for a testifier to convey knowledge to us. Footnote ⁸

A low trustworthiness assessment typically requires that the testifier provide more evidence than a more trustworthy testifier would need to provide. Suppose, for example, that someone had recently lied to you about whether it was raining outside or not. If they now come to you and assert that it's raining, you might reasonably require more evidence than just their word that it's raining (you might require corroborating testimony, or that they show you their wet hat, etc.). In contrast, if someone who had not recently deceived you about the weather came and told you that it was raining, you might believe them on their word alone.

A low plausibility assessment works similarly: when someone testifies to something implausible, their testimony is harder to believe – they will, at least usually, need to have more or stronger evidence than someone testifying to something more plausible. Thus, lower credibility assessments make it more difficult for testifiers to transmit knowledge. When we violate the constraints on legitimate influence between assessments of credibility, we wrongly deflate our overall credibility assessment. This, then, makes it more difficult for the testifier to transmit knowledge to us than it ought to be. This is our epistemic mistake.

6. From epistemic mistake to epistemic injustice

Of course, not all epistemic mistakes are epistemic injustices (Fricker Reference Fricker2007: 21–2). We regularly assign less trustworthiness to speakers than they deserve; this is always an epistemic mistake, but it is not always an epistemic injustice. These trustworthiness deflations rise to the level of injustice when the reason for the trustworthiness deflation is prejudice (Fricker Reference Fricker2007: 22). Similarly, there are many experiences for which we lack the conceptual tools necessary to understand those experiences. Insofar as we are unable to make sense of these experiences, we are worse off, epistemically, than we would be if we had those conceptual tools. But the absence of these tools only rises to the level of an injustice when the reason we lack these tools is systemic prejudice (Fricker Reference Fricker2007: 152).

So, does the illegitimate influence of an assessment of one ground of credibility on the assessment of the other ground ever rise to the level of an epistemic injustice and not merely a noteworthy epistemic mistake? I argue that it does.

We might think that this kind of illegitimate influence rises to the level of an epistemic injustice whenever the initial implausibility assessment or untrustworthiness assessment is the result of prejudice. In cases like this, unjustly low assessments of trustworthiness or plausibility do even more damage than they would have had they been reined in. Suppose, for instance, that a racist cop's unjustly low trustworthiness assessment of a Black male testifier is allowed to influence his assessment of the plausibility of that testifier's claim. This influence allows the cop's racism to doubly disadvantage the testifier – it not only unjustly affects the cop's assessment of his trustworthiness, but does further damage insofar as it affects his plausibility assessment. We might think this further damage constitutes an additional epistemic injustice.

Alternatively, we might think that this is a case that features one epistemic injustice (testimonial injustice) and another mere epistemic mistake. This mistake may compound the harms of testimonial injustice, but it does not count as an additional injustice.

I remain neutral here. I think either interpretation is plausible, so long as we acknowledge that, if it's not an additional epistemic injustice in cases like this, there's still an epistemic dynamic here worth taking note of. But I also think that, even if this is a case of this kind of epistemic injustice, it is not a central case, or an exemplar (though, I emphasize, this does not mean it is an unimportant epistemic misstep). Here's what I propose we consider to be the “central cases”: cases in which the available epistemic resources are structured so as to encourage the illegitimate influence of the assessment of one ground of credibility on the assessment of the other, and the epistemic resources are structured this way as a result of systemic prejudice. In these cases, the illegitimate influence can be said to be somehow caused by prejudice, and so rises from a mere epistemic mistake to a full-fledged epistemic injustice. In the next section, I argue that rape myths function in this way.

7. Case study: R v. Ghomeshi

Consider the following case:

The judge, William Horkins, cleared Ghomeshi because he believed the defense had cast adequate doubt on the victims’ credibility.

The three complainants were made to try their cases separately. This was significant; typically, the cases would have been tried together, and the similarities between the three women's allegations would have counted as evidence against Ghomeshi. Tried separately, this evidence was unavailable. The complainants were made to try their cases separately because the defense introduced text messages and Facebook messages exchanged between the three women as evidence of possible collusion (Kingston Reference Kingston2016). In the messages, the women discussed the impact Ghomeshi had on their lives and, in one, vowed to “sink the prick” because he sexually assaulted them (Kingston Reference Kingston2016).

These messages could only be interpreted as evidence of collusion if seen through a lens of distrust. Without the distortion of distrust, these messages might actually be considered evidence in the women's favor. This suggests that Horkins began the case with a low initial trustworthiness assessment.

Once the defense had the three cases tried separately, their strategy was to undermine the victims’ credibility. For the sake of space, I will focus on only one of the three women, Lucy DeCoutere, but all three women had their credibility attacked in similar ways.

One way the defense discredited DeCoutere was by focusing on the narrative structure of her claim. In his verdict, Horkins claimed that when DeCoutere recounted the assault, the details often shifted: sometimes she said Ghomeshi slapped her first, sometimes she said he choked her first. This, he said, made it hard to trust her account of the events.Footnote ⁹

Another way the defense discredited DeCoutere was by criticizing her for omitting information in her initial report to the police. DeCoutere did not initially reveal to the police that she had kissed Ghomeshi before he began to assault her, that she had kissed him goodbye on the night that he assaulted her, or that she later sent him flowers to thank him for taking her out – this information was all given later. DeCoutere claimed that she didn't think it was relevant; when she first talked to the police, she focused on the details of her assault, not on what had happened before or after (ONCJ 2016: 155). But Horkins claimed that he could not believe that “someone who was choked as a part of a sexual assault, would consider kissing sessions with the assailant both before and after the assault not worth mentioning when reporting the matter to the police” (ONCJ 2016: 155).Footnote ¹⁰ Further, he speculated that perhaps what motivated DeCoutere's “questionable conduct” was her role as an advocate for sexual assault victims. He admitted that he didn't have conclusive evidence for this, but maintained that it was still a “live question” (ONCJ 2016: 155).

Horkins’ assessment that DeCoutere's “shifting narrative” made her difficult to trust seems to be an instance of distrust being self-fulfilling. This minimal shift in details does not reasonably undermine DeCoutere's trustworthiness. This degree of variation seems to be well within the bounds of what is, in normal circumstances, understandable, and DeCoutere was testifying to an event that happened a decade ago – it would hardly be surprising if she couldn't remember whether Ghomeshi slapped her or choked her first, but did remember that he did both. However, if you approach her narrative through a lens of distrust – and so are prone to seeking out and fixating on inconsistencies, and overlooking explanations of those inconsistencies, as Jones claims – then the shifting details may be taken to undermine her trustworthiness.

Horkins’ lens of distrust, besides exacerbating his assessment of DeCoutere's untrustworthiness, also seems to color his assessment of the plausibility of her claim. He is explicit that he thinks it “unbelievable” that a victim of sexual assault would omit that she kissed her assailant before and after the assault, or that she sent him flowers afterward. His comment suggests that he thinks her behavior is out of line with what a “real” sexual assault victim would do. Importantly, his assessment came after DeCoutere explained that she didn't discuss kissing Ghomeshi before or after the assault or sending him flowers afterward when reporting the assault to the police because she (correctly) thought it was appropriate to focus only on the details of the assault, not on what occurred before or after. In failing to take this explanation seriously, Horkins seems to be allowing his distrust to color his assessment of the plausibility of her claim.

This is already evidence that Horkins allowed his plausibility and trustworthiness assessments to feed off of each other. However, most incriminating for Horkins is his speculation that perhaps what explains DeCoutere's “questionable conduct” is her role as a sexual assault advocate. Here, his assessment that her story is implausible is explicitly being allowed to influence his assessment of her trustworthiness – rather than investigating, on independent grounds, the plausibility of her claims, he simply believes that DeCoutere is lying instead.

This influence is illegitimate. First, Horkins’ initial assessment that her claim is implausible is unreasonable. DeCoutere has explained why she omitted those details when reporting to the police – she was reporting only those details relevant to the assault. Her omission of those details is not “astonishing”.

Further, and more importantly, Horkins’ reasoning seems to exhibit the pattern of runaway credibility deflation identified by Jones. He approached the case with distrust, as evidenced by his dismissal of potentially corroborating evidence. This distrust, plausibly, colored his assessments of the plausibility of DeCoutere's claims. Then, his assessments that her claims were implausible confirmed his initial assessment that DeCoutere was untrustworthy. This is exactly the pattern of runaway credibility deflation identified by Jones, and it makes it too difficult for DeCoutere to receive uptake.

This illegitimate reasoning is a central case of the kind of injustice that I have identified. Rape myth acceptance leaves people with an unjustly narrow conception of rape, which makes them likely to deem most rape allegations implausible because they fail to conform with their expectations of “real” rape. Rape myths also contend that women often lie about rape. Since this is a common belief, it coheres well with our background knowledge. Thus, when we hear an “astonishing” sexual assault case, and are searching for a way to make sense of it, it will be easier to accept that the victim is simply lying than it will be to accept that most rapes defy our distorted expectations. This means that rape myths encourage hearers to, upon hearing an astonishing case, suspect the veracity of the victim rather than seriously consider, on independent grounds, whether or not her claim is plausible. This kind of injustice goes beyond the wrongs done by hermeneutical and testimonial injustice – this injustice leaves some victims who come forward with a nearly insurmountable credibility deficit caused by the unjust interaction between assessments of trustworthiness and plausibility.

8. Hasty acceptance

In the case just discussed, the victims’ credibility was unjustly deflated because their assaults failed to align with the judge's expectations and he approached the testifiers with distrust. I now want to suggest that violations of the constraints identified can result in unjust credibility inflation, and so can explain what goes wrong in those cases where we believe an allegation on too little evidence.

A catastrophized conception of rape only works to make those cases we deem non-catastrophic difficult to conceptualize as rape. There are, however, cases that we readily recognize as catastrophic; these cases will not be rendered implausible by a catastrophized conception of rape. Whether a case is plausible will depend in large part on whether we are capable of seeing the accused as monstrous (recall: if the accused isn't monstrous, we may be led to an exonerating modus tollens inference). Many factors influence who we are capable of conceptualizing as a monster. Here I will restrict my consideration to the role race plays in structuring who we can picture as capable of rape.

The myth of the black rapist ascribes to black men irrepressible, animalistic sexual urges, thus making it seem like black men are more likely (or even very likely) to rape. This myth arose in the late 1870s as justification for ongoing lynchings (Davis Reference Davis1981: 186), and it has been difficult to shake. Many feminist scholars writing in the 1970s and 1980s, like Susan Brownmiller and Jean MacKellar, invoke the myth in their work on rape, suggesting that black men are more likely to rape than white men (Davis Reference Davis1981: 178). As I have discussed, Dylann Roof invoked the myth as he shot nine black men and women at the Emanuel African Methodist Episcopal Church (Bouie Reference Bouie2015). In February 2019, Liam Neeson discussed his reaction to discovering a close friend of his had been raped – upon hearing that she had been raped by a black man, he roamed the streets with a club, looking for any black man so that he could kill him. This reaction – the thought that any black man will do – only makes sense if all black men are seen as rapists, and so are all seen as equally guilty. In other words, Neeson saw this rape as having been committed by black men generically; he was attributing to black men, as a class, the essence of being a rapist.

The myth of the black rapist works to unjustly position black men as plausible rapists (and, so, plausibly monstrous). Because they are positioned as plausible rapists, allegations that they have raped someone can seem unjustly plausible. Moreover, black men are vulnerable to testimonial injustice – if, as Medina has argued, trustworthiness is comparative and contrastive, a trustworthiness deficit suffered by a black man can result in a relative credibility excess for women (Medina Reference Medina2013). So, white, cis, able-bodied women's allegations, in addition to seeming (potentially) unjustly plausible, may also seem more trustworthy than black men's denials. Since the white woman who accuses a black man has both an unjust testimonial advantage and benefits from an unjustly inflated plausibility assessment, I suggest that we may see the reverse of what was seen in the Horkins case – rather than having runaway credibility deflations, we may have runaway credibility inflations: the plausibility of the accusation may confirm the initial assessment that the white woman is a trustworthy testifier (relative to the black man, anyway), which may in turn reconfirm the plausibility of the allegation. This dynamic, then, might explain why we are so quick to believe some rape allegations – rather than investigating, on independent grounds, whether what is alleged is true, we allow our trustworthiness assessments and plausibility assessments to feed off of each other, resulting in unjust certainty that the allegation is true.