1 What Is Violence?

‘Violence’ is used to refer to many things. In the Western philosophical literature, it has primarily been used to refer either to intentional, excessive physical force or to the violation of morally considerable objects like persons, personal property, or non-human animals.² The main problem with the former is that it excludes some things we might want to call violence (like psychological or institutional violence).³ The main problem with the latter is that it includes too much, and by understanding violence as violation we run the risk of watering down the term’s meaning.⁴

In this project my goal is not to reinvent the wheel, nor is it to give a full-throated defense of either account of violence. Instead, my project is to adopt the second definition and then expand it in order to make sense of the claim that silencing can be an act of violence. In doing so I will ultimately adopt Vittorio Bufacchi’s definition of violence as violation of integrity. First, in order to illuminate both Bufacchi’s definition and my subsequent application, it will be helpful to review Newton Garver’s classic account of violence (from which Bufacchi borrows).

Garver argues that violence can be overt or covert, personal or institutional. What is common to each is not force but violation, and not just violation of any type of object, but the violation specifically of a person.⁵ Since “person” is a metaphysically robust concept, and for Garver violence is fundamentally about violating persons, violence itself should be understood to be similarly robust. Persons, after all, are more than their physical bodies; they also have at least some of the following: beliefs, desires, interests, self-awareness, the capacity for sensory experience, as well as the capacity to plan for and anticipate the future. Perhaps most important for Garver is that persons have a will; in having a mind with desires and interests, as well as the ability to plan for the future, persons are able to work to bring about for themselves one future rather than another.

On Garver’s view, violence can be done to persons in three ways. The first is by violating the body of a person; this is straightforward physical violence, where someone’s bodily integrity is undermined.⁶ The second is by violating the dignity of a person.⁷ This is best understood as a violation of another’s autonomy; that a person has essentially the capacity to make decisions and choose one option over another means that when another violates their ability to choose, they don’t merely undermine or thwart their choice, but what it is even to be a chooser in the first place. Finally, the third method of violence on Garver’s view is violating the freedom of a person to do what they would actually choose to do.

Bufacchi refines and moves beyond Garver’s view when he develops this definition:

An act of violence occurs when the integrity or unity of a subject (person or animal) or object (property) is being intentionally or unintentionally violated, as a result of an action or an omission. The violation may occur at the physical or psychological level, through physical or psychological means. A violation of integrity will usually result in the subject being harmed or injured, or the object being destroyed or damaged.⁸

Bufacchi argues that Garver’s analysis of violence is too exclusive, since it precludes the possibility of doing violence to non-persons (most notably, to non-human animals). That’s one reason why it is helpful to broaden the definition to focus specifically on violations of integrity.⁹ By “integrity” Bufacchi means to use the term “in a non-philosophical sense,” and to refer to “wholeness or intactness,”¹⁰ “‘unity,’ or the quality or state of being complete or undivided.”¹¹ Persons, like bridges and skyscrapers, have an integrity that can be undermined, causing them to fracture, become unstable, or to collapse altogether. And, since non-human animals and personal property have an integrity that can be degraded or destroyed, they can be the objects of violence as well.

I contend that Garver’s analysis is also too inclusive, in so far as he claims that thwarting the freedom of a person to act as they would is a type of violation. This casts the net too wide, since it would allow all efforts to prevent another from achieving their chosen ends to count as violence. For instance, if a parent secures a cabinet with a child-proof lock to prevent their toddler from getting into the toxic cleaning supplies, though they have thwarted their child’s ends, they have not done violence to them. That said, there is something to Garver’s third type of violation. In particular, it helps to make sense of the claim that threats are a form of violence, since what you do when you threaten someone is attempt to constrain their freedom and force them to comply with your will. It also helps to make sense of the claim that violence can be institutional. If some background social structure systematically thwarts someone’s choices over time it does violence to them. And, cutting now in the other direction and returning to the interpersonal, if a domineering parent routinely and habitually undermines their child’s decisions, not as they do when they prevent a toddler from drinking poison (which ultimately promotes the child’s freedom) but in a way that leaves them in a state of constant frustration, it makes sense to say that they do violence to their child.

This helps to illuminate a crucial insight to Garver’s project. Though Bufacchi is right to want to build a more inclusive definition than Garver (since it seems clear that violence can be done to non-persons), Bufacchi is also right to follow Garver in recognizing that there is a type of violence that can be done uniquely to persons. In such a case it is someone’s “integrity as a person that is infringed, since in the process of being violated one is reduced to a lesser being, in physical and/or psychological terms.” (Emphasis added.)¹²

Susan Brison develops a similar understanding of violence. Throughout her extraordinary book Aftermath: Violence and the Remaking of the Self, she explores in light of her own experience of violence what it is for a person to be “undone, demolished, shattered, and destroyed.”¹³ Like Garver, Brison argues that one method of causing such destruction is to undermine the autonomy of another. Since one characteristic of personhood is the capacity for autonomy, one method of violating persons is to undermine that capacity. When someone violates another’s ability to choose, they don’t merely thwart the person’s choice, but undermine what it is to be a chooser in the first place. That will often come about as a result of being routinely and systematically prevented from exercising one’s autonomy but can also be the result of isolated, traumatic actions.¹⁴

Because my task is to explore the claim that silencing can be an act of violence, and since I take persons to be the only types of objects that can be silenced, for the remainder of this project I will use “violence” to refer to the violation of the integrity of persons in particular. A person’s integrity can be violated both physically (by destroying someone’s body you violate not just their physical form but also their capacity to choose) and non-physically.¹⁵ I am here modifying the traditional distinction between physical and psychological violence; by contrasting the physical with the non-physical (rather than psychological) we are able to recognize types of violations of integrity someone might suffer that are not bodily, but also that are not straightforwardly psychological. Specifically, I want to consider epistemic violence, in which a person is violated in their capacity as a knower.

Physical and non-physical violence can be carried out both interpersonally (where one person violates the integrity of another) and institutionally (where some institution or social structure violates the integrity of a person or group of persons). Institutional violence can occur independently of particular actors, which is to say that it is a part of the fabric of society and is not necessarily intentional or aimed at by those who designed the laws, policies, or institutions in question.¹⁶ Understood in this way, social structures can cause violence, even if they leave no physical marks on their victims, and even if there are no clear perpetrators that can be held responsible for their harms.

Importantly, the account of violence on which I will rely is an outcome-based model of violence, which is to say that whether something is an act of violence is determined by whether its outcome includes the violation of the integrity of a person, rather than by determining in advance (by way of intention or how the act is performed) whether it so qualifies. In this way my approach aims to start from the standpoint of the victim, asking what happened to them, rather than the perpetrator, asking what they intended to do or how they intended to do it.¹⁷ In short, when a person or social structure violates the integrity of a person or group of persons, violence has been done.

At the same time, one might commit an action violently, even though the action does not itself qualify as violence. Bufacchi notes that it is helpful to distinguish the adverb “violently” from the noun “violence,” where the latter refers to what action was committed and the former refers to how the action was committed.¹⁸ It makes sense to say that someone performed an action violently even if they did not cause violence in so acting (as when someone with the flu coughs violently), just as it makes sense to say that they caused violence even if they did so via what would typically be taken to be non-violent means (as when someone accidentally poisons another by unknowingly serving them food to which they are seriously allergic). In short: if someone shoots a gun at another and misses they have acted violently though they have not committed an act of violence, since they did not violate the integrity of the person they were attempting to harm.

This distinction might seem strange; after all, actions like firing guns at people seem intrinsically violent, so why should we need to determine their outcomes in order to assess whether they are acts of violence? That strangeness is diminished when we note two things. First, though actions committed violently often cause violence they do not always do so, a fact that is easier to recognize when we consider attempted but unsuccessful acts more generally. Just as it makes sense to say that someone committed attempted (rather than successful) burglary, so too does it make sense to say that they committed attempted (rather than successful) violence. Second, part of the reason why the outcome-based account of violence may appear problematic is that it might seem to suggest that the actor who tries but fails to do violence (though they do act violently) does not act wrongly. But, just as we can hold someone morally responsible for attempted burglary, so too can we hold someone responsible for attempted violence. The same is true for acts that do not aim at violence but could foreseeably cause it. In other words, the strangeness of distinguishing between acting violently and committing an act of violence is diminished in light of the fact that we already frequently draw that distinction (when distinguishing between attempted and successful acts), as well as when we recognize that whether (and to what degree) someone should be held morally responsible is a different question from whether they committed an act of violence or performed an act violently.

Putting these pieces together it becomes clear that the world contains considerably more violence than is typically recognized.¹⁹ One might be troubled that this analysis of violence is much more inclusive than ordinary uses of the term imply. As I said at the outset, my project here is not to argue for this analysis of violence but instead to adopt it in order to make sense of the claim that silencing can be an act of violence. I do believe, however, that this more expansive analysis is helpful because it focuses on the underlying moral issue that we should care about in many cases of harm, wrongdoing, or injustice. Killing, physical assault, and rape all involve a violation of a person’s bodily integrity. Threats, coercion, exploitation, being Otherized or treated with disrespect, can all involve violation of a person’s will, of their dignity or sense of self, or of their social status or position. And, as I will argue, silencing can involve violations of a person’s epistemic capacities, thereby undermining or diminishing their ability to serve as a giver and receiver of knowledge, and short-circuiting their ability to interpret the world in a way that is meaningful and that fits with their experience. In short, part of what is valuable about the analysis of violence as violation of the integrity of a person is that it elegantly identifies an important and unifying feature of many of the serious harms that can be inflicted on others. It also is consistent with Brison’s testimony as a survivor of violence as well as her analysis as a philosopher. It seems both analytically and practically useful, then, to tie a term as normatively laden as “violence” to that underlying feature. If, however, you object to this definition, you may simply substitute “violation of the integrity of a person” for “violence” throughout the rest of this chapter, since I will take the latter to refer to the former.

2 What Is Silencing?

Employing Garver’s taxonomy of violence, silencing can occur interpersonally or institutionally, overtly or covertly. Here are some examples of the way that silencing plays out in all four senses (though this list is far from exhaustive).²⁰

Overt Interpersonal – If one person physically restrains another (putting their hand over the other’s mouth), or disables their access to the internet or social media, or constantly talks over them such that they are never able to effectively communicate their view, they have silenced the other overtly.²¹

Covert Interpersonal – If one person in conversation with another constantly rolls their eyes, responds patronizingly, or fails to pay attention to the speaker, who then stops speaking rather than suffer such continued disrespect, they have silenced the other covertly.

Overt Institutional – If the state passes a law that prevents its citizens from communicating with each other – if it blocks social media or shuts down the free press – the state has silenced its people overtly by thwarting their ability to communicate in the first place. Or, if the state passes a law that allows only a particular social group to vote, all other groups have been overtly silenced institutionally.

Covert Institutional – If the state enacts various types of voter suppression, making it more difficult and costly for citizens to vote in an election and thereby discouraging communication of their political will, then the state has silenced its people covertly. Or, if a dominant ideology or features of a culture train members of a particular social group not to attempt to communicate or even to have a view in the first place, members of that group have been covertly, institutionally silenced.²²

In short, I take silencing to occur when someone is prevented from communicating.²³ Such an outcome can be brought about more or less overtly by both individuals and institutions or social structures. As with violence, we should adopt an outcome-based account of silencing: we determine whether someone has been silenced by determining whether they have been prevented from communicating. (Consider a case where the state tries (but fails) to shut down the press or where one person tries unsuccessfully to talk over another at the dinner table; such cases would be instances of attempted but unsuccessful overt institutional and interpersonal silencing, respectively.)

This definition is more inclusive than those generally defended in the literature. Though Rae Langton grants that one way to silence someone is to “prevent them from speaking at all,”²⁴ she is concerned primarily with the ways in which someone might be silenced even though they are able to express themselves. Jennifer Hornsby and Langton later defend what has come to be the go-to definition of silencing, which says that someone is silenced when their illocutionary intention in communicating is prevented systematically from securing uptake by their intended audience.²⁵ Mary Kate McGowan argues that silencing also happens not only when someone’s illocutionary intention fails to secure uptake, but when a speaker’s sincerity – whether they mean what they are saying – in performing a particular speech act fails to secure uptake.²⁶ Kristie Dotson outlines two particular types of silencing, what she calls testimonial quieting and testimonial smothering, both owing to pernicious ignorance on the part of the intended audience. Testimonial quieting “occurs when an audience fails to identify a speaker as a knower.”²⁷ Testimonial smothering occurs when “the speaker perceives one’s immediate audience as unwilling or unable to gain the appropriate uptake of proffered testimony.”²⁸ My definition is compatible with each of these accounts – it recognizes all of them as silencing – but it also includes additional forms of silencing that these authors might omit, since, for instance, my view does not entail that silencing be grounded in pernicious ignorance, that it be systematic, or that it be focused in particular on the intended audience’s behavior (though many instances of silencing might entail all three conditions).

Have I cast the net too wide? One reason for accepting a more inclusive definition of silencing is that it better syncs up with ordinary usage. Consider Case 1 with which I began. In the current political climate in the United States, talk of silencing on college campuses is common. When student protesters disrupt or cause public lectures to be canceled, those speakers (and those commenting on the events) have often claimed that they were silenced. And, in many cases, preventing the speaker in question from communicating what students anticipate will be morally repugnant views is exactly what the protestors aim to accomplish. This is overt, interpersonal silencing. Or, consider Case 2, in which students in a college class make sexist jokes or appeal to sexist stereotypes that thereby make it difficult for women in the class to take part. This is covert, interpersonal silencing.²⁹ Next, consider Case 3, in which a college blocks the use of a particular smartphone app on its network on the grounds that it was being used to communicate homophobic beliefs. This is overt, institutional silencing. Finally, consider Case 4, in which a company fires its employee for communicating racist or sexist beliefs. In this case the company does not silence the employee; that they were able to communicate their views is what led to them being fired. Instead, the company sends a message that others who express such views will also be sanctioned. This is covert, institutional silencing.

Another reason to adopt a more inclusive definition is not only that it better tracks contemporary, ordinary usage, but also that it better captures the underlying component that runs throughout many of the analyses that can be found in the literature. Hornsby and Langton, for instance, are fundamentally concerned with the disruption of communication, and focus on how misogynist ideology (and the prevalence of pornography in particular) undermines the way in which women are able specifically to communicate that they do not consent to sex.³⁰ In contrast with Hornsby and Langton, Rebecca Kukla explores what she calls discursive injustice, in which (moving beyond a focus on silencing) the performative force of a speaker’s speech act is distorted, owing to the background context in which they are attempting to communicate.³¹ What is common throughout the literature is that the attempt to communicate is thwarted, either because the person to whom the communication is aimed fails to do their part (either intentionally or unintentionally), or because the background context is such that communication itself is difficult or impossible. Consider again Case 4: in firing the employee the company communicates to others that the background conditions within which they might express themselves are such that, should they express certain types of views that are inconsistent with the company’s values, they will lose their jobs. Though these background conditions are not as covert as those detailed by Hornsby and Langton and by Kukla, they help set the terms within which communication can take place (and so also create the circumstances in which silencing can occur).

For both of these reasons, then, I contend that we should accept a broad definition of “silencing.”³² Like violence, we should adopt an outcome-based account of silencing, in which we determine whether someone was silenced in light of whether they were prevented from communicating. It then makes sense to reach the judgment that people were silenced in all four of the cases I named at the outset. The question I now want to engage is whether such silencing should ever be considered violence.

3 Is Silencing an Act of Violence?

Miranda Fricker argues that essential to one’s identity as a person is one’s identity as a knower and that to suffer from epistemic injustice is to be “wronged in one’s capacity as a knower.”³³ But what does it mean to be wronged in that way? On Fricker’s view, there are two types of epistemic injustice: testimonial and hermeneutical injustice. Someone suffers testimonial injustice when their testimony is not assigned the degree of credibility that it warrants or when they are not recognized as a credible source of knowledge in virtue of their social location (owing to what she calls “identity prejudice”).³⁴ So, if women are taken to be unreliable or to lack credibility with regard to certain types of knowledge claims, when a particular woman attempts to provide that type of knowledge, her efforts will likely fail to secure uptake. And, such lower levels of credibility need not only be assigned in virtue of particular types of knowledge; if women as a social group are seen to be unreliable or unknowledgeable, then women’s testimony will tend to fail to secure uptake across the board. Silencing, then, can be one form of testimonial injustice, in that the person who is silenced is prevented from communicating on the grounds that they are not credible sources of knowledge.

The harm here is not merely the offense of not being heard, of not being recognized as someone whose testimony is (or even could be) veridical, but the cumulative, developmental effects that such failure of recognition can have over time. In being effectively excluded from the community of knowers, one is prevented from “steadying the mind.”³⁵ As essentially social beings, we constantly check our beliefs against the beliefs of others, confirming or disconfirming them in light of those of others whom we encounter.³⁶ As Lisa Guenther argues in her work on the phenomenology of solitary confinement, we constantly rely on others to help us to solidify our beliefs that the world is the way that we understand it to be.³⁷ Over time, when prevented from engaging in such confirmation, our faith in our own understanding of the world can erode, leaving us less sure of ourselves, our sensory perceptions, methods of reasoning, and memory, and so less able to go about making knowledge claims in the first place. In other words, each of those capacities can atrophy without use or can become distorted with misuse, and so epistemic isolation can leave one diminished in their capacity as a knower.³⁸

Testimonial injustice’s effects can be magnified when it leads to or is accompanied by hermeneutical injustice, which occurs when “some significant area of one’s social experience [is] obscured from collective understanding.”³⁹ This is born from having been hermeneutically marginalized, which is the exclusion of a social group from participating in and contributing to the collective understanding.⁴⁰ If you and others who share your social location are prevented systematically from contributing to the collective understanding, you’re likely to encounter many instances of mismatch between your perception of the world and the majority (or at least epistemically dominant) group’s perception of the world.⁴¹ Silencing can contribute to or cause hermeneutical injustice, in that the person who is silenced is prevented from communicating their own experience and interpretation of the world. Their hermeneutical marginalization then is likely to compound, making it more and more difficult for them to be able to contribute to the collective understanding.

Again, as with testimonial injustice, the harm is not merely offense at marginalization and exclusion (which is both appropriate and considerable) but the way in which, cumulatively and over time, suffering from hermeneutical injustice would cause you to lose trust in your own epistemic capacities. Someone who consistently experiences hermeneutical injustice might begin to feel “crazy,” to doubt their own perceptive and cognitive capacities, and to wonder whether they are able to form beliefs in a way that syncs up with the world as it actually is.⁴²

In short, suffering from epistemic injustice can make you less able to exercise your epistemic capacities across the board, including your imaginative and interpretive capacities, sensory capacities, and capacity to reason. Since those capacities are constitutive of your being a knower, and being a knower is essential to being a person, being subjected to epistemic injustice can have the effect of diminishing you as a person. Applying the account of violence I adopted earlier, it then makes sense to recognize that epistemic injustice is a form of violence when it has such effects. Since violence is the violation of the integrity of persons, and epistemic harms can violate an essential part of someone’s personhood, when someone is subjected to epistemic injustice (and it has the effect over time of violating their integrity) then they have been subjected to epistemic violence.⁴³ And, when silencing is the method by which epistemic injustice is enacted or perpetuated, silencing becomes an act of violence.

Though epistemic injustice is a form of violence – specifically epistemic violence – when it results in the violation of the integrity of persons, since not all instances of epistemic injustice have the effect of violating the integrity of persons, not all epistemic injustice is a form of violence. Whether epistemic injustice is a form of violence is determined by whether any person’s integrity is violated as a result, which depends on the context, the way epistemic injustice plays out within that context, and the personal constitution of those who are affected by it.

Dotson has developed a different account of epistemic violence. She argues that successful communication relies on both the speaker and the hearer doing their parts, and that “to communicate we all need an audience willing and capable of hearing us.” (Emphasis in original.)⁴⁴ She then goes on to define epistemic violence as: “a failure of an audience to communicatively reciprocate, either intentionally or unintentionally, in linguistic exchanges owing to pernicious ignorance. Pernicious ignorance is a reliable ignorance or a counterfactual incompetence that, in a given context, is harmful.”⁴⁵ Note that my view is more inclusive than Dotson’s in two ways. First, my view does not require that the audience suffer from pernicious ignorance. Second, and more important, my view is not focused in particular on some failure of the audience to hear, but instead allows that someone might be silenced in various ways that do not depend on the intended audience’s actions at all. Though I agree completely with Dotson that both ignorance and a failure to communicatively reciprocate will often cause epistemic violence, neither are necessary on my view. Instead, in adopting an outcome-based account of violence we determine whether violence has been done by asking whether a person’s integrity has been violated, which can occur independently of prejudice on the part of the hearer or on the hearer’s failure to reciprocate.

Just as Dotson’s view of epistemic violence is too restrictive (since it requires that epistemic violence be grounded in pernicious ignorance), so too is Fricker’s account of epistemic injustice too restrictive (since it requires that epistemic injustice be grounded in identity prejudice).⁴⁶ As I have argued elsewhere, it is possible to cause the negative effects of epistemic injustice absent any identity prejudice.⁴⁷ So, I contend that just as we should adopt an outcome-based account of violence and silencing, so too should we adopt an outcome-based account of epistemic injustice. Or, barring that revision to Fricker’s account, we should at least recognize that acts of silencing that are not grounded in identity prejudice can cause the same effects for those who are silenced as epistemic injustice has on its victims. Since those effects violate the epistemic integrity of those who are subjected to them, silencing can sometimes do epistemic violence to someone, regardless of whether it adheres to a strict definition of epistemic injustice.

Epistemic violence can be just as damaging as many forms of physical or psychological violence. It need not be inflicted intentionally (and indeed often is not). And, it can happen independently of any particular actor, but can instead be born from the everyday processes of normal life in an otherwise well-meaning society.⁴⁸ Return to Case 2: if the professor does nothing to prevent women in their class from being excluded from the community of knowers, then even though the professor doesn’t believe that women ought to be attributed less credibility than men, women in the class might still suffer testimonial injustice as a result. In the same way hermeneutical injustice is also possible. In a case of an extremely powerful hegemonic ideology, it might be literally unimaginable for someone to conceive of the world in some other way. They might manage to endorse and recommend that ideology to others through countless actions and thereby perpetuate it without intending to.⁴⁹ If that ideology is harmful and leaves those whose experience of the world fails to sync up with it worse off in a way that undermines their capacity as knowers, they do violence to others even though they not only don’t intend to, but cannot even imagine that their actions might do so.

Note that, according to the outcome-based account of violence I have been developing, in a situation where someone is silenced but they do not suffer deleterious effects to their epistemic capacities, the silencing they experience is not an act of violence. This implication is one that some might find troubling. If someone is very confident and secure in their epistemic capacities and is not undermined by being silenced, then even if they are silenced, since their integrity has not been violated, violence has not been done to them. Return to Case 1, in which a speaker is prevented from giving a public talk at a college campus. The speaker might be very confident and sure of their epistemic capacities; though they might feel frustrated that they were prevented from speaking (either because their invitation was rescinded or because protests made their talk difficult to hear) they do not suffer a loss of faith in themselves. In that case they have not suffered violation, and so though they were silenced, that silencing was not an act of violence. That does not mean that the silencing to which they were subjected was not wrongful or in other ways harmful – it might very well be. Furthermore, it’s also important to recognize that, even though the speaker might not suffer epistemic violence in such a case, others who witness the silencing might not be so confident and secure in their epistemic capacities. If those capacities are diminished as a result of the speaker being silenced, those audience members might suffer collateral epistemic violence. Regardless of whether such collateral violence occurs, even though (on my view) some actions that we might commonly call violence turn out not to be (in so far as no actual violation of integrity occurs) that doesn’t mean that attempted violence should not be blamed, nor does it mean that acts that are unintentionally or accidentally acts of violence require blame: those remain open questions that require additional argumentation and context. Similarly, consider a case in which one person attempts, but fails, to silence another. That they have neither silenced the other nor done epistemic violence to them does not mean that they have not acted wrongly; that also remains an open question. My point here is that for silencing to be an act of violence, it must result in a violation of integrity. The question of whether silencing someone in a way that causes epistemic violence is wrongful is the one to which I will now turn.

4 When to Silence Others

There are many good reasons to silence someone. Consider, for instance, standard examples of what are commonly taken to be acceptable legal limitations of free speech, like the fact that it is illegal to disingenuously shout “fire” in a crowded place, or for someone who knows classified information to share it with the press, or for doctors to disclose sensitive information about their patients. Teachers and professors also silence their students in various ways. They prevent their students from communicating many things in class discussion by helping them stay on topic if they start to wander and by telling some talkative students they need to be quiet to create space for less talkative students to contribute. Finally, parents and guardians limit their children’s time on social media or their phones, thereby preventing them from communicating freely.

All three sets of examples illustrate that it is uncontroversial to think that it can be morally appropriate for someone to be silenced, either because doing so promotes the greater good (preventing someone from shouting “fire,” sharing classified information), because doing so prevents the violation of others’ rights (patients’ rights to privacy, quiet students’ rights to take part in class discussion, and all students’ rights to have that discussion be a productive and effective learning environment), or because it is in the best interests of the person who is silenced (limiting a child’s access to social media). So, though silencing sounds bad (and often is bad) it is often positively good (as well as totally mundane).

I suggest instead that it is not silencing simpliciter to which we should object, but silencing that does epistemic violence to others. There are additional, instrumental reasons to object to silencing, like concerns about an overly interventionist state or the ways in which social and scientific knowledge might be hindered as a result, but I won’t explore them here. Instead, my purpose is to focus just on the claim that silencing can be an act of epistemic violence. Why worry about that silencing in particular? For the same reason that we should worry about violence in all its forms. If violence is the violation of a person’s integrity, and persons are intrinsically valuable, then all violence is prima facie bad; if we lived in an entirely just and morally perfect world then there would be, at the least, no nonconsensual violence.⁵⁰ But, we do not live in such a world and the prima facie reason not to commit violence is trumped by other sometimes stronger factors, like the right to defend yourself from violence and the obligation to protect others from violence.⁵¹

In appealing to both the right of self-defense and the obligation to defend another, I am assuming that, in such cases, the only way to protect yourself and others is by committing acts of violence. But, in those cases, though violence is bad, it is still permissible or obligatory. Since silencing can, in some circumstances, violate a person’s integrity (undermining their epistemic capacities in a way that undermines their personhood), and since some types of speech have the effect of silencing others, in those circumstances where you must silence someone in order to protect yourself or others, you ought to even if doing so will be an act of epistemic violence.

Consider again Case 2: given background sexist ideologies, it is all too easy for women to feel as if they don’t belong in a college environment in general, or in a specific class or discipline in particular. Imagine that, as a result of systematic exposure to such ideologies (and the jokes and comments in Case 2 that help to perpetuate them), a student comes to doubt herself deeply and to lose trust in her epistemic capacities, such that her epistemic integrity is diminished and rendered unstable.⁵² In short, she becomes less of a knower, and since being a knower is an essential part of being a person, her integrity as a person is undermined. That student suffers, in short, from epistemic violence by way of testimonial injustice as a result of having been silenced by those who use such language. The professor has an obligation to try to prevent such violence; is it permissible for them to do so by silencing others who would use such language in class, even if doing so would constitute epistemic violence for those silenced in the long run?

Things become more complicated when we recognize that it’s not just particular terms (like sexist language) that can have such an effect; some views, even when expressed in ways that appear to be polite or follow norms of decorum, are themselves likely to be silencing. Consider, for instance, the sexist stereotype that women are bad at math; even if a particular math teacher is not sexist and conducts class in a gender-neutral way, it’s likely that at least some women in the class will still be silenced by stereotype threat and the norm that says that “people like them” don’t have anything to contribute to conversations like those held in class.⁵³ If others in the class make jokes that tread on that stereotype, or circulate news articles that claim to have discovered why women are bad at math, should the professor prevent the expression or discussion of the view?

I hope to have demonstrated that while a great deal of silencing is unobjectionable, one primary reason why silencing becomes problematic is when it is an act of epistemic violence. But, I have also taken for granted that it is sometimes permissible to use violence to prevent other violence, either directed at yourself or another, at least when there are no other options available. At the same time, as I have argued in defending an outcome-based account of violence, we often don’t know in advance when an action will do violence to another. And, when we recognize that silencing can sometimes be an act of violence (even if unintentionally so) how does someone know when they ought to silence another? I don’t intend here to offer a one-size fits all answer. Instead, what follows is a list of general factors that ought to be considered when trying to decide whether silencing that does violence to another would be justified in a particular case:

1. –Someone’s speech would likely cause epistemic violence or be otherwise harmful to others;
2. –The speaker would be culpable for the epistemic violence or harm they cause (because they should know better);
3. –Silencing would be beneficial to the speaker (in that you prevent them from committing culpable epistemic violence against another);
4. –Silencing someone better serves an underlying commitment to communication, better promotes the flourishing of the epistemic community, or enhances (rather than undermines) speech in general;
5. –Silencing prevents the communication of speech that would promote injustice by contributing to unjust ideologies that are likely to do violence to subordinated social groups;
6. –Silencing promotes hermeneutical justice by allowing those who might otherwise be hermeneutically marginalized to contribute to the collective understanding, thereby promoting a more complete, less partial and distorted collective understanding of the world;⁵⁴
7. –Silencing would do the least amount of harm or violate the fewest duties and rights of all those affected.

I do not claim that this list is exhaustive, nor do I claim that each one of these factors would always outweigh all other considerations in all contexts. Furthermore, I do not claim that whether any of these factors will obtain is always knowable; indeed, as is very often the case in trying to act rightly, you run a moral risk in which your efforts might end up doing more harm than good. But, since it is the case that refraining from action is itself action, opting out altogether is not an option. In light of that fact, it is my hope that in trying to answer the thorny and painful questions exemplified by the four cases with which I began, these desiderata are morally and politically useful.

The advocate for the unfettered right to free speech might reply that we could easily avoid the difficulty of such questions by being willing to risk causing epistemic violence and pursuing truth wherever it goes, independent of such moral or political consequences. This is the wrong approach to take, in part because as feminist epistemologists have spent decades demonstrating, there is no neutral place to stand in choosing which questions to ask or which lines of discussion to pursue. What that means is that in choosing to throw open the door and allow all conversations to take place while recognizing that some conversations will do epistemic violence to members of particular social groups, you in fact take not a neutral position, but one that perpetuates existing injustice. In short, if you grant that silencing can be an act of violence but you claim that we should ignore moral and political considerations like those I named above, then you do take a side, despite claiming to adopt a position of neutrality in which all positions should be free to see the light of day.

5 Conclusion

In this chapter I have defended the claim that silencing can be an act of epistemic violence. Though I believe that it can and that it is one that is very common, we should not despair. In adopting an outcome-based account of violence we saw that silencing becomes an act of violence when someone’s epistemic agency is violated. Since such violation often (though not always) occurs as a result of systematic and repetitive instances of epistemic harm over time, I remain hopeful about the ways in which we can shore up each other’s epistemic capacities and help each other to become less vulnerable to violation. Indeed, in the long run it should be the goal of those working to promote social justice to bring about a world in which everyone is deeply confident in their own epistemic abilities and where no ideologies would tread on anyone’s social location in a way that would leave anyone vulnerable to epistemic violence. That is not to say that we ought to aim for the type of individualistic imperviousness that my critic might claim everyone ought to pull themselves up by their epistemic bootstraps to achieve. It is instead to note that it is precisely because we are fundamentally social beings who are susceptible to the influence of others that we should work to create the conditions where everyone is able to trust themselves in a way that would render all disagreement free of violence. Whether we could ever achieve such a world remains an open question. In “the unjust meantime,”⁵⁵ some views ought to be silenced in order to avoid foreclosing that possibility.