Epistemic rights violations and epistemic injustice

I offer a detailed discussion of the connections Lani Watson posits in her book The Right to Know between violations of a person’s epistemic rights on the one hand and the distinctively epistemic forms of injustice influentially discussed in the work of Miranda Fricker on the other. I argue that if we specify the content of the relevant epistemic rights (and the corresponding duties of others) carefully enough, it becomes plausible that there is an even tighter relationship between violations of these rights and various epistemic injustices than Watson claims. I close by tabling a speculative suggestion on the significance of this conclusion.


Introduction
Epistemologists have been slow to unlearn the myth of the single, autonomous inquirer in science, philosophy, and more generally, but the sheer extent of our epistemic dependence on others has gradually become clearer. However, with this epistemic dependence on others comes vulnerability, and that vulnerability is one that individual people, institutions, companies, advertisers, politicians and governments, and various others exploit all the time. This is not a new issue, but it has reached new levels in the information/internet age, in which a relatively small group of people and institutions exercise a lot of power over the flow of information-and misinformation. Given this, we need protection; in The Right to Know, Lani Watson argues that we need epistemic rights. 1 Watson's book is structured around five main and interrelated questions. What are epistemic rights? Who has them? When are epistemic rights violated? What harms 1 3 29 Page 2 of 14 are done when peoples' epistemic rights are violated? Why do we need epistemic rights? I'm going to focus here on the fourth question, about epistemic rights violations, and only discuss Watson's answers to the other questions to the extent that doing so gives necessary background for engaging with the fourth. In particular, I will examine what Watson says about the relationship between epistemic rights violations and epistemic injustice in Miranda Fricker's sense (Fricker 2007): wrongs done to one in one's capacity as a knower, or more generally, in one's capacity as an epistemic agent. In Section 2, I will lay out some relevant details of Watson's discussion of epistemic rights. Section 3 offers a preliminary look at the connections Watson draws between violations of such rights and different forms of epistemic injustice, then the following two sections examine these connections in more detail, arguing in each case that refinements of Watson's proposals are plausible; Section 4 focuses on what Fricker calls testimonial injustice, while Section 5 turns to what she calls hermeneutical injustice. Section 6 concludes and briefly speculates on the significance of the connections Watson draws for our understanding of epistemic injustice.

Epistemic rights
What are epistemic rights? As Watson concedes, this is not an expression we hear very often, but the title of her book gives us one familiar idiom in which they are frequently discussed in both political and legal discourse and more colloquially: "the right to know". This also suggests one shape that epistemic rights can take, namely, having the right to have access to or be furnished with certain information or knowledge. That is too narrow, though. We care about epistemic goods other than knowledge: evidence, justification, understanding, wisdom, accuracy, and so on. 2 Moreover, if we have epistemic rights, they can take various different forms. Suppose I have a right to know the outcome of a medical test that has been performed on me. I also, it seems, have the right not to have my doctor tell other people the outcome without my consent, and I may have the right not to have to disclose the outcome to others. I may also have the right not to know the outcome of the test. 3 Watson is well aware that the locution "the right to know" really only picks out one possible kind of epistemic right, and so she develops a much richer taxonomy of rights; we do not need to get into the details of that here, but the general point will be relevant when we think about the epistemic rights that might be associated with the proffering of testimony.
2 Some epistemologists hold that at least some of these epistemic goods are in effect just other ways of talking about knowledge and the roles it plays: see McGlynn 2014 for critical discussion. 3 Whether one has such rights may depend on what the test is for, and whether there are other parties with their own epistemic rights; for example, if the test is for a sexually transmitted disease, it may be that not learning the outcome would be reckless on my part, and there could be other affected parties who have the right to know the outcome even though the test was performed on me. Who has what epistemic rights in a given scenario may be a complex matter and may need to be assessed on a case-by-case basis.
Watson starts by specifying what a "right" is, in the sense that she wants to claim we have (or should have) epistemic rights. Following the American jurist Wesley Newcomb Hohfeld, Watson defines a right as "a complex entitlement that provides justification for the performance and prohibition of actions and omissions" (2021, 3). There are a couple of distinctive aspects of Hohfeld's conception of rights, which Watson explicitly takes on board along with this definition. First, rights are not really entitlements to things, but rather to "the performance and prohibition of actions and omissions." We talk of a right to food, for example, but on this conception this is really shorthand for a right to have people do (or not do) certain things that contribute to people having food. Watson thinks that it can often be harmless to speak as if one has the right to certain goods, including epistemic goods such as knowledge, but sometimes we need to be more careful.
The second distinctive aspect of Hohfeld's approach is that there are, in a sense, different kinds of rights: privileges, claims, powers, and immunities. Here is how Watson explains privileges, rights, and immunities: Privileges remove any restrictions there might be on me doing something but they do not oblige me to do it nor place a duty on anyone else to help or hinder me. Powers allow me to waive a claim that I have, in effect removing the duties placed on others by my claim. Immunities prevent others from taking away or altering the claims that I have and the duties that attach to them.' (Watson 2021, p. 4) As these characterisations suggest, there are important relationships between rights and duties-however, it is the notion of a claim right that is core to those relationships, and so Hohfeld and Watson pick claim rights out as special within this scheme: Hohfeld highlighted the significance of the relationship between claims and duties. He noted that for every claim there is a correlative duty. In other words, if I have a claim, then someone else has a duty towards me either to act or not to act in a certain way. […] Claims determine what duties people have with respect to others' rights.' (Watson 2021, p. 4) The other three kinds of rights lack this immediate connection to duties; for example, as characterised above, privileges expressly do not involve correlative duties for others. Given this, Hohfeld and Watson hold that only claim rights are rights in the truest sense; the other kinds are rights in an attenuated sense. For this reason, when Watson talks about epistemic rights in her book, she is almost always focused on epistemic claim rights, together with their correlative epistemic duties (Watson 2021, p. 5). The function of such rights is to protect rights holders (Watson 2021, p. 7), and this fits with Watson's motivations for proposing that we need epistemic rights, as noted in my introduction.
On Watson's view, rights are justified by, or have their basis in, the "final value" of rights holders, rather than on instrumental grounds; it is in virtue of them being "persons", in a morally loaded sense. It is helpful to contrast this view with the kind of account Watson is opposing here. One might think that right holders have rights because those rights enable them to lead happier lives, so that rights would be instrumentally grounded in the way they can contribute to holders getting to something else that is valuable (Watson 2021, p. 11). Watson is saying, in contrast, that the justification for rights holders having rights is just the kinds of beings they are, regardless of whether having those rights contributes to them acquiring or producing other things of value or not. 4 It is not enough to violate one's epistemic duties to act (or omit to act) in ways that do not fulfil or accord with those duties. This is merely to infringe on a person's rights, and such infringements may be justified. Suppose I grab your mobile phone out of your hand and throw it in a river; this clearly infringes on your rights to have and use your phone. But if I know you are about to call in a bomb-hoax to the local hospital, this infringement might be completely justified. 5 To violate someone's rights is to unjustifiably disregard "any duties resulting from that right, to perform, or not to perform, certain actions" (Watson 2021, p. 47).
So far we have been examining theses about rights in general that Watson takes from Hohfeld. But turning to our real quarry, we can ask, what makes a right an epistemic right? Watson answers that these are rights to epistemic goods, conceived broadly so as to include not only knowledge, but also "truth, belief, justification, understanding, wisdom, information, misinformation and ignorance" (Watson 2021, p. 14, see also p. 13). 6 A violation of a person's epistemic rights, then, involves others unjustifiably disregarding any duties resulting from that person's rights to epistemic goods.
There is much more to be said about rights in general, and epistemic rights in particular, much of which is explored in detail and with both insight and nuance in Watson's book. However, we have enough of her framework on the table now to turn to the issues I want to focus on.

Epistemic rights and epistemic injustice
In a paper published a few years before her book, Watson writes I argue (in work in preparation) that epistemic justice and injustice are a function epistemic rights, meaning that all forms of epistemic justice involve respect for epistemic rights and all forms of epistemic injustice involve the violation of epistemic rights. (Watson 2018, p. 101 n11) There is a relatively strong claim made in this passage: that all violations of epistemic rights involve epistemic injustice, and all forms of epistemic injustice involve epistemic rights violations. Watson does not say much in defence of this claim here, and she goes on in the remainder of this passage to mention just one form of epistemic injustice, "informational injustice", involving unjust distributions of information and opportunities to access information; following Fricker, this sort of unjust distribution of access to epistemic goods has generally been treated as rather peripheral in the literature on epistemic injustice. 7 Presumably the "work in preparation" mentioned in this quote is Watson's book, but in that book she offers what seems like a much more qualified claim about the overlap between epistemic rights violations and epistemic injustice: It is useful, I think, to conceive of epistemic justice and injustice as a function of epistemic rights. By this I mean that, when epistemic rights are respected, epistemic justice will typically prevail. Conversely, when epistemic rights are violated, epistemic injustice will typically be the result. Thus, epistemic justice typically involves respect for epistemic rights and epistemic injustice typically involves the violation of epistemic rights. I say typically because, while the relationship between rights and justice is a close one, it is not one of conceptual necessity. In other words, rights can be violated without the occurrence of injustice and injustice can occur without the violation of rights. Nonetheless, epistemic rights violations can often be identified in cases of epistemic injustice. (Watson 2021, p. 69) In the book, Watson is also more engaged with the two varieties of epistemic injustice discussed by Fricker. These are testimonial injustice, which involves a person receiving less credibility than they ought to from their audience due to that audience's prejudices, and hermeneutical injustice, which involves unjust obstacles to members of marginalised social groups making their experiences intelligible to themselves or to others due to appropriate hermeneutical resources (such as concepts or linguistic expressions) being unavailable.
It is fairly clear that epistemic rights violations will involve what Watson calls "informational injustice", and vice versa, as she claims in her 2018 paper. If we all have the right to know certain things, but there are barriers put up such that my right is respected and yours is violated, that seems to result in an unfair distribution of knowledge (particularly if the barriers are systematic, discriminating against an entire social group that you belong to). Likewise, if there is an unjust distribution of epistemic goods, it is plausible that this will involve violations to the epistemic rights of those receiving less than their due share.
It is less clear what, if any, the relationship is between epistemic rights violations and the other two species of epistemic injustice discussed by Watson. Moreover, Watson softens her claims between her 2018 paper and the book, only asserting that epistemic rights violations and epistemic injustice "typically" or "often" go hand in hand, and so we need to investigate whether these qualifications help to make the claimed connections defensible, and whether they are needed. In the next section, I will examine what Watson says about epistemic rights violations and testimonial injustice, and in the sequel I will do the same for hermeneutical injustice.

Testimonial injustice
Suppose that a speaker has been subjected to a testimonial injustice; she has prejudicially received less credibility from her audience than she is due. Watson suggests that the audience had a duty to "give credit to the speaker", but they unjustifiably disregarded that duty, and so "the speaker's epistemic right to be believed without prejudice is violated" (Watson 2021, p. 69). This might sound plausible enough on the face of it, but so stated, I do not think that speakers have any such epistemic right. An epistemic right to be believed without prejudice is an epistemic right to be believed, and speakers do not in general have any such claims on their audiences. One may have a right to be able to express one's views (though of course there are well known and fierce debates about how absolute or qualified such a right is), but that falls well short of a right to have one's audience come to believe one's views, so expressed. To say this much is not to say a speaker never has the right to be believed when testifying. Take one of the central examples Fricker presents as a paradigm case of testimonial injustice. 8 In To Kill a Mockingbird, as Fricker interprets the story, Tom Robinson receives a credibility deficit when testifying about what happened on the day he stands accused of raping a young white woman, Mayella Ewell, where this is due to racist prejudices held by the jury and others. In his closing statement, Robinson's lawyer, Atticus Finch, almost pleads with the jury to do their duty, adding (as Harper Lee depicts the scene, almost to himself, so that the rest of the courtroom and the reader are left a little unsure what he says) "In the name of God, believe him" (Lee 1960, p. 227). Fricker takes Finch to be "trying to impress upon the jury that they have a duty to believe Tom Robinson" (Fricker 2007, p. 25), and it seems plausible enough that the jury do indeed have such a duty. It is less plausible that this exemplifies any general duty that audiences have to believe speakers, though. One reason this kind of example seems special is the particular dynamics of the courtroom setting and the jury's role within the legal process; I will come to a second reason unto think this case may be unrepresentative in a moment. 9 Perhaps the duty hearers have is not to "give credit to the speaker" or to uphold a right to be believed, as Watson's wording suggests, but more carefully put, to give the speaker the degree of credit they are due. The degree of credit a speaker is due may not be high, and in particular, it may not be enough to warrant believing what they say; there is, on this proposal, no general right to be believed, but only a right to receive a degree of credibility that one deserves, not one distorted by prejudice. In cases like Tom Robinson, in which the only reason for not believing his testimony is prejudice, perhaps we can and should say there is a right to be believed, but again this emerges as a special case. 10 If that is right, can we identify epistemic rights that are plausibly violated in all cases of testimonial injustice, not just in cases that share particular special features?
We can start by looking at how Fricker specifies what a credibility deficit involves: Epistemological nuance aside, the hearer's obligation is obvious: she must match the level of credibility she attributes to her interlocutor to the evidence that he is offering the truth. (Fricker 2007, p. 19) 11 If we can move freely between talk of obligations and talk of duties, the proposal here is that the hearer has a duty to match her credibility judgment to the evidence that the speaker is telling the truth. If she fails to do this by prejudicially giving the speaker less credibility than the evidence indicates he ought to have, she violates one of his epistemic rights.
This cannot be the whole story, however, and we need to complicate this account of the connection between testimonial injustice and epistemic rights violations. After all, it is part of Fricker's picture that we can go some way towards repairing credibility deficits (and more generally, we can work on cultivating the virtue of testimonial justice, becoming disposed to make such repairs, and perhaps eventually even becoming less disposed to rely on the prejudicial stereotypes in currency in our community in the first place). 12 Such repairs involve revising the degree of credibility one awards a speaker upwards, to roughly where it would have been had prejudice not interfered in one's original credibility judgment. As this rather vague description suggests, the making of such revisions is not an exact science for Fricker; so long as one ends up in the right ballpark, she seems to suggest, one has done enough (Fricker 2007, pp. 91-2). In principle, then, the revised degree of credibility one awards may still count as a deficit if it fails to match where it should be given one's evidence (and in can in principle even be a credibility excess, if one overcompensates for one's prejudices). Presumably the idea is that if one does a sufficiently good job of counterbalancing the influence of one's prejudices, a residual credibility deficit no longer counts as prejudicial even if there is still a slight mismatch with the degree of credibility one ought to give, given the evidence one has that the speaker is telling the truth.
This suggests that failing to match one's credibility judgment about a speaker to the evidence that they are telling the truth need not violate, or even impinge on, that speaker's epistemic rights; one might engage in imperfect but adequate repair work, ending up with a final credibility judgment that is close enough. Put differently, a hearer's duty is to avoid prejudice as far as they are able and to mitigate its impact when it has a foothold; speakers have the corresponding epistemic rights that are violated when hearers unjustifiably fail to carry out these duties.
If this is on the right track, it suggests a more intimate connection between testimonial injustice and epistemic rights violations than Watson was willing to assert in her book, once we specify the content of the relevant rights. It is not just that testimonial injustice often or even typically involves the violation of a subject's epistemic rights, but rather than it always does. Crucially, though, the epistemic rights violated may be less demanding that the right to be believed that Watson herself focuses on.

Hermeneutical injustice
Here is what Watson writes about hermeneutical injustice and epistemic rights violations, continuing on from the claims about testimonial injustice quoted in the previous section: In cases of hermeneutical injustice, a complex set of duties attaches to groups and institutions within a community, correlated with the epistemic rights of its members to share in the practices that determine meaning for the community. Hermeneutical injustice often arises when such duties are unjustifiably disregarded and the associated epistemic rights are violated. Both these forms of epistemic injustice [testimonial and hermeneutical] typically involve the violation of epistemic rights. (Watson 2021, p. 69) The idea here is that members of a community have epistemic rights to contribute to meaning-determining practices; the practices that shape that community's collective hermeneutical-interpretive-resources. Institutions and groups within that community have corresponding duties; Watson does not specify what these are, but they would seem likely to be duties not to interfere with people's participation in the relevant meaning-determining practices, and perhaps to facilitate their participation. 13 Watson then claims that when such duties are unjustifiably disregarded, hermeneutical injustice "often arises", and so she concludes that hermeneutical injustice, like testimonial injustice, typically involves the violation of epistemic rights.
There are a number of issues raised by this brief passage. First, Watson writes here as if it is individual epistemic agents who have the relevant epistemic rights, while "groups and institutions" have the corresponding complex set of epistemic duties. But we might wonder whether we should really be talking about the rights of individual members of an epistemic community to participate in its meaning-making practices here. After all, one might think that hermeneutical marginalisation-exclusion from and marginalisation within these practices-happens primarily at the level of social groups, rather than that of individuals. That I am personally unable to participate in those practices does not seem to have much significance for epistemic injustice, if I belong to a social group which is on the whole well represented and able to exercise power within my community's meaning-making practices, and so whose experiences tend to be readily rendered intelligible by my community's collective hermeneutical resources. So perhaps Watson's point in the passage quoted is better put in terms of the epistemic rights of social groups within a community rather than individual members. 14 Even with that change in mind, there is room to wonder precisely what Watson is claiming in the passage above; what does it mean to say that hermeneutical injustice "often arises" when people's epistemic rights to contribute to hermeneutical resources are violated? One interpretation is that the phenomenon of hermeneutical marginalisation involves the violation of an individual or a group's epistemic right to contribute to their community's meaning-making practices, and (on Fricker's picture at least) hermeneutical injustices are a consequence of such marginalisation. 15 That is to say, the background condition for hermeneutical injustice, hermeneutical marginalisation, involves widespread and systematic violations of the epistemic rights of the social group that the victim of that injustice belongs to.
Watson may also mean to suggest that particular instances of hermeneutical injustice are themselves instances of violations of the epistemic rights of individuals, not merely that the kinds of rights violations involved in hermeneutical marginalisation are present in the background. An example will help to bring out what I have in mind, and why we might think this further claim is plausible. Let us take Fricker's central example of hermeneutical injustice, Carmita Wood struggling to make sense of her experiences of being, as we would now put it, sexually harassed by a professor at Cornell, and to make her experiences intelligible to others (Fricker 2007, pp. 149-151). On Fricker's interpretation of Wood's predicament, she (and everyone else at that time) lacked the concept of sexual harassment, and so Wood lacked a piece of knowledge about herself, as well as the expressive resources to describe precisely the behaviour the professor was inflicting on her and the ways it was harmful to her. It is plausible that this example, so understood, involves layers of epistemic rights violations; Wood belonged to a social group that was prejudicially excluded from their society's meaning-making practices, and as a consequence she lacked pieces of knowledge about herself that she ought to have had, and which it was very much in her interests to have. The hermeneutical injustice she faced itself seems to have been a violation of her right to know, on top of those involved in being a member of a hermeneutically marginalised group.
As has become clearer in the subsequent literature on the topic, there are many cases of hermeneutical injustice where this additional layer is missing. These are cases in which, unlike Carmita Wood (as Fricker understands the example), the epistemic agent has the conceptual resources to understand their social experiences just fine, and even to communicate about them with other members of the marginalised group they are a part of, but faces obstacles when trying to make their experiences more widely intelligible due to the necessary hermeneutical resources not having gained sufficiently widespread currency: due to these concepts and expressions not having become part of the community's shared stock of hermeneutical resources. 16 In examples of this sort, the speaker does not lack any piece of self-knowledge or self-understanding; rather, they struggle to make their experiences "communicatively intelligible" to particular audiences. 17 So there is no piece of knowledge they have a right to but lack. But as with testimonial injustice, members of hermeneutically marginalised groups may have additional epistemic rights associated with the reception of their testimony, perhaps corresponding to the description Fricker gives of what the virtue of hermeneutical justice calls for. She suggests that this involves being attentive to the possibility that someone coming across as unintelligible may be struggling with a hermeneutical injustice, trying to help them overcome the barriers to communicating their experiences when investing that much time and care into an exchange is possible, and keeping an open mind when it is not possible (Fricker 2007, pp. 171-172). As I did above in the case of testimonial injustice, we can identify the epistemic rights of speakers as the correlates of this complex set of duties on the part of audiences. 18 The picture that is emerging is one on which hermeneutical injustice involves layers of epistemic rights violations, some of which are downstream from others. At the root are the epistemic rights violations involved in hermeneutical marginalisation: violations of the rights of social groups and their members to contribute to and shape their society's communal hermeneuetical resources. As a result, members of hermeneutically marginalised groups will be vulnerable to facing further violations of their epistemic rights: violations of their right to know, and in particular their right to certain important items of self-knowledge, or violations of their right not to face unnecessary further resistance or indifference when struggling to make themselves intelligible, or-as, perhaps, in the case of Carmita Wood-both.

Epistemic (in)justice as a function of epistemic rights
In light of the discussion in the previous two sections, I want to suggest that Watson is right when she suggests, in the passages quoted above, that we should "conceive of epistemic justice and injustice as a function of epistemic rights" (Watson 2021, p. 69), but that she does not need to qualify the links she draws between epistemic rights violations in the manner she does in her book. 19 We need not rest content with saying that epistemic rights violations "typically" or "often" give rise to one or both of the two varieties of epistemic injustice that Fricker identifies, or vice versa. Rather, I am proposing that once we get clearer on the content of the epistemic rights violated in cases of different kinds of epistemic injustice, we can maintain the kind of tighter links that Watson seems to have found appealing in her earlier 2018 paper; we can have a picture on which there are complex and interrelated epistemic rights associated with contributing to one's society's communal hermeneutical resources, and with sharing testimony to an audience, and different kinds of epistemic injustices involve those epistemic rights being violated in different ways. One wrinkle here is that Watson holds that in general rights violations and injustices can be dissociated in both directions: that there can be rights violations that are not injustices and injustices that do not involve violations of rights (Watson 2021, p. 69). This general point is one reason she posits a qualified link between epistemic injustices and rights violations in her book. Unfortunately, Watson does not offer any examples, and so it is difficult to know to what extent this point really does undermine the kind of connection I am proposing here.
But supposing that we can maintain an intimate link between epistemic injustice and epistemic rights violations, what is the significance of this? I want to close with a speculative proposal, one which I cannot explore here but which I think merits further consideration; we can use this relationship between epistemic injustice and epistemic rights violations to explain why the former counts as a form of injusticewhy in particular the notion of injustice is applicable here, rather than some other morally loaded notion. This is a question that has often been rather neglected in the literature. 20 According to the proposal I have in mind, the reason that the wrongs that Fricker discusses count as forms of injustice is that they involve violations of epistemic rights, rights which (on Watson's view, one might recall from Section 2) people have in virtue of being beings with final value: persons. This might not strike us as particularly informative as it stands, but it may prove illuminating and important to locate epistemic injustice within the more general and better-understood category of rights violations-in particular, I am suggesting that it may illuminate the character of the wrongs involved in epistemic injustice in a way that enables us to better see why they should be thought of as injustices.
I have argued here that cases of epistemic injustice, in Fricker's sense, invariably involve violations of a person's epistemic rights, but not for the converse claim. There can be unjustified violations of a person's epistemic rights which are not epistemic injustices in Fricker's sense, as when my doctor fails to impart certain medical information about myself to me due to incompetence or being overworked, rather than any kind of prejudice. I may be epistemically harmed by the doctor's omission, but there does not seem to be any sense in which I am wronged or insulted in my capacity as an epistemic agent. So the category of epistemic rights violations is broader than epistemic injustice, as Fricker conceives of it. Despite this, I am suggesting that we explain why the phenomena Fricker focuses on count as injustices in virtue of their 20 This point has been stressed to me a number of times by Catarina Dutilh Novaes. A recent exception is Byskov (2020), which exacts an answer to this question from Fricker's own discussion, and argues that it requires supplementation. I lack space to discuss Byskov's proposal, or to compare and contrast it with what I am suggesting here, but I do so in future work. An anonymous referee points out that there is plenty of literature that tries to specify the kind of inherent wrong involved in epistemic injustice: in Fricker's terms, these are accounts of the primary harms of epistemic injustice and of testimonial injustice in particular: see, for example, Fricker (2007, chapters 6 and 7), Pohlhaus (2012), Congdon (2017), Giladi (2017), Hawley (2017), Goetze (2018), and McGlynn (2020. I do not take these proposals to be attempts to answer the question in the text, however. They are accounts of the inherent, non-contingent epistemic, and moral wrong involved in different varieties of epistemic injustice, and such an account should explain the way in which a particular kind of epistemic injustice involves a wrong done to a person in their capacity as a knower or epistemic agent-that is, how it meets Fricker's general characterisation of epistemic injustice (Fricker 2007, p. 1). For instance, Fricker holds that the primary harm of testimonial injustice is that the speaker is epistemically objectified, thereby wronged in their capacity as a giver of knowledge, and that this explains the sense in which testimonial injustice involves a wrong done to a person in their capacity as a knower. However, this does not explain why we should think of the wrong here as a form of injustice in particular; for Fricker, the reason we should think of these wrongs as injustices is not because they involve a speaker being objectified, but seems to be rather because of the way that testimonial injustice (and, importantly, hermeneutical injustice) involves prejudice and discrimination. We might also note that discussions of primary harms treat testimonial and hermeneutical injustice separately (with the exception of Giladi, who tries to offer a more or less unified account of the primary harms of both). In contrast, the question in the text-what makes epistemic injustices injustices-seems to invite a unified answer. The answer floated in the text is that epistemic injustices involve violations of people's rights, specifically their epistemic rights; that answer seems compatible with any of the main proposals concerning primary harms (or indeed, with scepticism that we can identify primary harms of these varieties of epistemic injustice at all). relationship to that broader, more general category of unjustified violations of epistemic rights. Does any particular significance still attach to epistemic injustice, as Fricker understands it, or does the proposal under consideration completely subsume Fricker's category within the more general category of unjustified epistemic rights violations? I like to hope that the former would still be true even on the proposal I am floating-that there would remain reasons to pick out and focus on injustices involving a person being wronged in their capacity as an epistemic agent-but only a much more careful investigation than I have offered here can settle the matter. 21