Identifying the Right Perlocutionary Failure

Perlocutionary silencing bears similarities to silencing more generally. Established accounts reveal that silencing harms speakers by illegitimately preventing speech acts their conversational participation entitles them to perform. From this characterization of silencing, we can establish the following argumentative framework to show that some failed perlocutionary acts silence speakers. If for some set of perlocutionary acts, conversational participation alone generates a defeasible entitlement to perlocutionary success, then any interlocutor behavior illegitimately preventing those acts silences the speaker. So, if we identify perlocutionary acts where conversational participation defeasibly entitles speakers to perlocutionary success and interlocutor behavior illegitimately prevents that success, we identify instances of perlocutionary silencing. This section establishes that illegitimately preventing a perlocutionary act that conversational participation entitles the speaker to perform deserves the label “perlocutionary silencing.”

Silencing theory identifies a linguistic harm beyond failed perlocution. Ishani Maitra and Mary Kate McGowan, for example, identify silencing as systematic communicative interference (Maitra Reference Maitra2004; Reference Maitra2009; McGowan Reference McGowan2009; Reference McGowan and Mikkola2017). They argue that silencing harms speakers by preventing them from fulfilling their communicative intentions when they could reasonably expect their intentions to be fulfilled. If conversational participants speak literally to an attentive audience who understands their words, speakers can assume their speech acts will succeed (Maitra Reference Maitra2004, 193). Nevertheless, the speaker's intended speech act might fail because the hearer may not believe the speaker to be authoritative or sincere, preventing the hearer from recognizing the speaker's intention (McGowan Reference McGowan and Mikkola2017). Essentially, silencing arises because something illegitimately prevents full conversational participation. So, speakers cannot achieve what their conversational participation minimally authorizes.

Initially, perlocutionary silencing appears problematic. Perlocutionary success requires satisfying nonlinguistic conditions that conversational participation does not guarantee. Perlocutionary acts are actions speakers perform by making utterances. Specifically, J. L. Austin characterized perlocutionary acts as the utterance's effect on the hearers’ or speaker's feelings, thoughts, or action (Austin Reference Austin1962, 101). In what follows, we shall include as perlocutionary acts the effects, intended or not, from having made an utterance, including actions and events resulting from changes in feelings, thoughts, or actions. Consequently, perlocutionary success depends on the utterance's consequences. Unlike communicative acts, perlocutionary acts lack the same guarantee of success.Footnote ¹ Speakers cannot always expect perlocutionary success from conversational participation among linguistically competent participants alone. Whether a speaker can expect perlocutionary success often depends on beliefs about the addressee rather than conversational participation. For instance, someone may tell a joke hoping to make their audience laugh, but whether the speaker can expect perlocutionary success depends, among other things, on what the speaker knows about the audience's sense of humor, whether the audience has heard the joke before, and whether the joke was told at an appropriate time. Similarly, if I invite friends to a dinner party, expecting they will attend, my expectation depends on my beliefs regarding whether they are free that night, whether they want to attend, and whether they feel comfortable eating the food I prepare, among other considerations. But each consideration goes beyond linguistic and conversational norms. Since perlocutionary acts depend on extralinguistic effects, and silencing interrupts speech, perlocutionary silencing appears misguided. Successful perlocution requires satisfying nonlinguistic conditions, and such failure hardly qualifies as silencing. Nothing in the conversation entitles the speaker to these conditions being satisfied.

Relatedly, failed perlocution only occasionally harms speakers. Consider assertion and a corresponding perlocutionary goal, persuading. Persuasion may routinely fail. If one person urges another to accept their assertion, that perlocutionary goal would fail if the interlocutor introduced a counterexample. The counterexample frustrates the perlocutionary goal but does not silence the speaker. Moreover, if I invite friends to a party, but they cannot attend, my perlocutionary goal fails. I may justifiably expect my friends will attend, but successfully inviting them, alone, does not justify my expectation. So, perlocutionary frustration is too broad a category for silencing.Footnote ²

Thus we see three related problems. First, perlocutionary acts are consequences of speech rather than speech itself. Second, perlocutionary success depends on extralinguistic factors, which conversational participation does not guarantee. Third, failed perlocution is too broad for silencing. Consequently, mere perlocutionary failure, that is, perlocutionary frustration, does not qualify as silencing.

Rather than objections, these worries offer a starting point for analyzing perlocutionary silencing. Rae Langton's original analysis of silencing was too broad (Langton Reference Langton1993). Yet theorists narrowed illocutionary disablement to failed communicative acts. This offers an approach: focus the concept on the specific harm we intend to identify. Accordingly, we identify a linguistic harm where the speaker's perlocutionary goal fails. Thus, we focus on interlocutor behavior causing perlocutionary frustration, constituting linguistic harms. Specifically, something interrupts the speaker's intention, and the interruption constitutes a linguistic harm. We can further focus on consequences of speech that conversational participation might guarantee or demand.

Examples reveal these features. When speakers assert, they may have various perlocutionary goals, but successful assertion does not guarantee every goal. Some failed goals harm speakers because they prevent them from fulfilling conversational obligations. When speakers assert, a challenge requires they defend or retract their assertion (see Brandom Reference Brandom1994). If the interlocutor challenges the speakers’ supporting assertions before the speakers defend their previous assertions, the interlocutor prevents the speakers from defending their assertions. Since the interlocutor prevents the speakers from fulfilling their dialectical obligations, we shall call the interlocutor a recalcitrant questioner.

Consider the following dialogue:

In this example, the interlocutor challenges the speaker to prevent them from adequately defending their assertions. However, the recalcitrant questioner never stops the speaker from defending their assertions by preventing further assertions. Instead, the recalcitrant questioner generates additional obligations. In this example, the recalcitrant questioner moves the topic away from the original assertion. The speaker made a claim about universal basic income but ended up defending claims about depression. If the dialogue continued, the recalcitrant questioner could prevent returning to the original topic or cause the speaker to retract their assertion based on an irrelevant challenge. Thus, the speaker may never fulfill their obligation to justify their assertions.Footnote ⁴

Here, asserting is the illocutionary act, but justifying is perlocutionary. Justifying an assertion is perlocutionary because speakers intend for their justifying assertions to bring about conversational participants’ acceptance as common ground the previous assertion's content. Defending an assertion is a perlocutionary dialectical move in the sense that by making an assertion with the appropriate inferential and epistemic relations to the previous assertion, the participants accept the previously asserted content as common ground. Speakers have defeasible reasons to expect that as long as they satisfy the appropriate inferential and epistemic relations, which may be contextually determined, their previous assertions will be accepted as common ground. Speakers are obligated to achieve this goal assuming they can satisfy the inferential and epistemic conditions for defending their assertions. Speakers’ expectations are further grounded in the assumption that their interlocutor is competent and cooperative. There is no guarantee the asserted content will be accepted as common ground, but the speaker must make moves to bring that content into the common ground or retract the assertion. By continuing to ask questions, the interlocutor prevents the speaker from possibly achieving that perlocutionary goal so long as an undefeated objection to the original assertion remains.

Recalcitrant questioners harm speakers as linguistic agents because they prevent them from fulfilling their conversational roles. Indeed, such interlocutors prevent speakers from fulfilling their dialectical obligation: defending their assertions. Recalcitrant questioners silence speakers despite letting them assert. Here we identify a perlocutionary harm but also a linguistic harm: speakers are prevented from reaching their dialectical goals despite performing their intended illocutionary acts. Speakers cannot do what their conversational participation minimally authorizes. Moreover, recalcitrant questioners prevent a speech act that the conversation requires. So, the recalcitrant-questioner example exemplifies two things we seek: (1) it involves a perlocutionary act, and (2) participation in the conversation entitles the speaker to reasonably expect perlocutionary success. Here, the speaker reasonably expects to make the necessary moves to defend their assertion because the conversation requires they defend their assertion. So, the interlocutor harms the speaker as a speaker trying to fulfill their dialectical responsibility. The recalcitrant questioner exemplifies a perlocutionary-level linguistic harm, and the harm depends on conversational participation rather than extralinguistic features.

The recalcitrant questioner exemplifies perlocutionary silencing because the interlocutor prevents a perlocutionary act that conversational participation entitles the speaker to perform. The example, however, utilizes assertion's idiosyncrasies. The recalcitrant questioner exploits assertion's dialectical norms. Perlocutionary silencing may arise across discursive contexts. So we shall benefit from another example.

Students often try to change classroom discussions, and instructors might dislike altering their course plans. But sometimes a teacher perlocutionarily silences a student by preventing them from influencing the discussion's direction. Here, to facilitate the example, we stipulate that an identity prejudice prevents changing a topic.

Suppose a class discusses Frankenstein. A female student wants to explore Frankenstein's motivation. The teacher says, “We'll return to that,” but never intends to because he believes it is not worth discussing with that student because she is female.Footnote ⁵

Here, we stipulate the instructor thinks that the student, being female, lacks the sophisticated knowledge to recognize Frankenstein's motivation's textual significance. Consequently, the instructor believes discussing Frankenstein's motivation with that student would end fruitlessly. Had the student been male, or had the teacher lacked his identity prejudice, the student's perlocutionary goal would have succeeded. Since the student intended changing the discourse topic as her perlocutionary goal, the example qualifies as perlocutionary frustration. The teacher recognized the recommendation as a recommendation, hence his response. So, the student's illocutionary act succeeded, that is, she successfully made a recommendation. Unfortunately, the instructor did not follow her recommendation.

The instructor silences the student because he prevents her from performing a perlocutionary act that her class participation entitles her to make. As a class participant, the student actively participates in the pedagogical conversation. Her illocutionary success illustrates her participation. Within the class context, various discussions would help the instructor achieve his pedagogical goals regarding the Frankenstein text, including discussing Frankenstein's motivation. The student recognized Frankenstein's motivation's significance but wanted to better understand it. Moreover, since the student is a class participant, she is defeasibly entitled to influence the conversation to enhance her learning. The student could reasonably expect perlocutionary success, but the instructor prevents a perlocutionary act that the conversation entitles the speaker to expect will succeed. This example reveals a further example, where conversational participation entitles a speaker to perlocutionary success, but an interlocutor's behavior illegitimately undermines success.

The Frankenstein example differs importantly from similar examples involving illocutionary silencing. Consider a meeting comprising mostly men, including a male manager running the meeting. In the meeting, a woman speaks, but no one responds. Moments later, a man repeats what the woman said. This time, the manager congratulates the man for such a good idea. The woman was silenced.Footnote ⁶ But, important for our current discussion, the woman's illocutionary intention failed. She did not convince her colleagues because they never credited her attempted assertion as an assertion. This example differs from the Frankenstein example where the illocutionary act succeeded. The teacher's willingness to consider the student's recommendation reveals the difference. Since the teacher responded to the recommendation, the student successfully recommended changing the discussion, but the teacher disingenuously dismissed her recommendation, preventing the student's perlocutionary goal. The teacher willingly engaging the student's recommended topic change reveals an important difference between the two cases.

Analyzing Perlocutionary Silencing

Thus far we have revealed that conversational participation defeasibly entitles speakers to perlocutionary success for some perlocutionary acts. Moreover, interlocutors sometimes illegitimately prevent those acts, perlocutionarily silencing speakers. This parallels illocutionary silencing. However, appropriately characterizing perlocutionary silencing reveals the parallel is not merely failed speech acts that speakers’ conversational participation entitles them to perform. Perlocutionary silencing parallels illocutionary silencing because it constitutes a significant moral harm by frustrating legitimate human interests. This section further argues that we should deem a class of perlocutionary failures perlocutionary silencing because of these harms. Such perlocutionary failures deserve the label “perlocutionary silencing” rather than merely “perlocutionary frustration.” I begin by characterizing perlocutionary silencing. Further reflection on the harm reveals the class of perlocutionary failures worth considering.

Locutionary and illocutionary silencing share a common feature: speakers cannot fulfill a linguistic intention. Silencing stymies discursive participation by interrupting speech acts that speakers reasonably expect will succeed because their conversational participation entitles them to perform these acts. Silencing prevents discursive moves that conversational participation authorizes. Yet, in the recalcitrant-questioner example, the speaker successfully performs locutionary and illocutionary moves. So perlocutionary silencing must differ. The relevant perlocutionary acts influence the conversation. Unlike the invitation example, the salient examples involve speech-related perlocutionary goals—perlocutionary goals changing the conversation. Accordingly, examples reveal a speech-related harm despite perlocutionary acts not being essentially linguistic. The recalcitrant questioner intentionally prevents the speaker from defending their assertion, derailing the conversation. In the Frankenstein example, the teacher rejects the recommendation because of an identity prejudice. Speakers experience unjustified harms as speakers attempting to influence the conversation.

Having identified an unjustified linguistic harm at the perlocutionary level, we can articulate perlocutionary silencing. A speaker is perlocutionarily silenced when they are illegitimately deprived of perlocutionary influence on the conversation in which they participate.

Certainly, more must be said regarding what qualifies as “perlocutionary influence on a conversation,” and to this we shall return. But first, we shall benefit from further examining perlocutionary silencing's moral dimensions. To show that the class of perlocutionary failures we identified as perlocutionary silencing is morally problematic, we shall consider two things: (1) perlocutionary influence on a conversation constitutes a legitimate human interest, and (2) interrupting perlocutionary influence frustrates a legitimate human interest.

Perlocutionary silencing is morally problematic because it sets back a central human interest. Mari Mikkola, following Joel Feinberg and David Archard, argues that moral wrongs indefensibly set back interests central to human function—interests necessary for minimal well-being (Mikkola Reference Mikkola2016, 145; see Feinberg Reference Feinberg1984; Archard Reference Archard2007). These interests involve setting a life plan and defining oneself, and they ultimately contribute to overall welfare. When someone violates one's minimal well-being, that violation constitutes a harm because it hinders the basic realization of that person's value. Anything that illegitimately interrupts realizing those interests, or illegitimately prevents acknowledging someone's value, constitutes a moral injury. Perlocutionary silencing frustrates such interests by interrupting our ability to fully function as human agents, for conversational moves constitute legitimate human interests.

Legitimate human interests define who we are, and we organize our lives around them. These interests give lives meaning and factor into achieving other interests (Mikkola Reference Mikkola2016, 165). Legitimate human interests enable humans to lead autonomous lives, allowing us to determine a good life and how to pursue it. They include continued life and health but also social interests. For example, human social capacities give our lives meaning when we act on those capacities. Important human interests include engaging in meaningful social relationships, including communication.

Meaningful communication includes directing conversations, which forms a legitimate human interest for at least two reasons. First, directing conversations is one way we realize our social capacities. When we influence conversations, we exercise our agency using speech. Second, it fosters other interests, making it central to our humanity. We see this most clearly when speakers direct conversations to their needs. These changes can range from directing the conversation to the speaker's immediate needs to how we should order society to achieve life goals, and from verbally defending oneself to clarifying an issue the speaker considers important. In these examples, speakers must direct the conversation to achieve their goals and, in some instances, even discover how to achieve goals.Footnote ⁷

Directing a conversation takes on a moral dimension because interlocutors can illegitimately interfere with this legitimate human interest. Within a conversation, participants essentially invite each other to use their linguistic agency, which entitles participants to discursive moves. Simply put, when individuals participate in conversations, each participant has a prima facie entitlement—and so a defeasible expectation—to influence the conversation's direction. Likewise, participants have defeasible reasons to let interlocutors make discursive moves. Although overriding reasons may exist, illegitimately preventing speakers’ conversational moves is unjustified. When interlocutors illegitimately prevent speakers’ conversational moves, they wrong the speakers. In other words, perlocutionary silencing constitutes a significant moral harm by frustrating a legitimate human interest, further paralleling illocutionary silencing.

The above account establishes perlocutionary silencing as morally problematic. Thereby we discovered that the harm arises because linguistic agency and conversational participation entitles speakers to certain discursive moves. This reveals what “perlocutionary influence on a conversation” means, but we can further clarify this concept. Perlocutionary effects influence conversations, especially conversational direction. A speaker might change the conversation by mentioning a passing car. If a speaker says something that makes a camel appear, the speaker will likely change the conversation. Yet interrupting such conversational moves would not silence speakers, for these speakers change the conversation through secondary perlocutionary influence. In the first, the speaker changes the conversation by drawing the hearer's attention to the passing car. If the conversation changes, it depends on whether the speaker can draw the hearer's attention to the car and how likely the hearer finds the passing car worth discussing. It does not directly depend on conversational norms. Likewise, if an utterance makes a camel appear, the camel will likely consume the hearer's attention, but any discursive change would be a secondary effect dependent upon the perlocutionary goal of making the camel appear. In both examples, the speakers make objects salient, but saliency influences interpretation while not determining discursive direction. Within a conversation, participants’ backgrounds and the surroundings render various objects salient, but these objects influence the conversation only when a participant refers to them.Footnote ⁸ What is important here is what the conversation defeasibly guarantees. Perlocutionary influences on conversations comprise discursive moves, such as the speaker attempting a change in subject, status, or conversational role.

Since perlocutionary conversational influence constitutes a legitimate human interest, and illegitimately interfering with these human interests generates a morally significant harm, we identify a class of perlocutionary frustrations as perlocutionary silencing. Silencing does not merely interrupt speech acts. Rather, silencing harms speakers by illegitimately preventing speech acts that conversational participation entitles speakers to make. Since speakers can reasonably expect some perlocutionary acts will succeed from conversational participation alone, illegitimately interrupting those acts constitutes a significant moral harm paralleling silencing more generally.

Legitimately Preventing Conversational Moves

The preceding account frames perlocutionary silencing in terms of illegitimately depriving speakers of conversational moves their participation defeasibly entitles them to. So any account of perlocutionary silencing must say something about these defeasible entitlements. Moreover, a thorough account must include when such entitlements can be overridden. To that, we now turn. This section reasserts that a speaker's entitlement arises from conversational participation and establishes conditions that would legitimately override speakers’ entitlements.

When a speaker is entitled to perlocutionary success depends on context. Apart from conversational exchange, norms governing conversations do not arise. Thus the sort of exchange determines the specific conversational norms, that is, participants’ obligations and entitlements. For example, friendly banter generates different norms from those in a committee meeting covering student retention. Despite the moral significance of how best to serve first-generation college students, without an additional reason, a participant in the banter conversation should not expect to change the conversation to first-generation college students, whereas a committee member can expect a successful change in topic. Accordingly, a speaker's entitlement to perlocutionary success does not depend on morally weighted reasons.Footnote ⁹ Thus, the speaker's legitimate expectation depends on the speaker's conversational participation rather than the content's significance.

Content matters, however, when assessing a speaker's conversation-shaping move with regard to conversation topic and antecedent speech acts. Again, the relevant entitlements depend on the conversation, and overriding reasons can depend on antecedent speech acts. In the banter and committee-meeting examples, participation entitles participants to recommend a topic change. However, in the former, the conversational goal, namely banter, justifies not following the recommendation, whereas in the latter, retention of first-generation students is germane, so the recommendation should be followed. Yet the recommendation can be overridden, for example, if the speaker was inattentive when the committee earlier discussed challenges that first-generation college students face. Here we see how content can matter based on antecedent speech acts within the conversation.

Providing a general framework for when overriding a speaker's entitlement qualifies as legitimate involves challenges because every attempt to influence a conversation is unique. An interlocutor's refusal to follow a speaker's attempt to guide the conversation counts as silencing only when that refusal is illegitimate. So we must say something about what counts as legitimate and illegitimate refusals. Conversational moves happen against a large backdrop. When questions arise about whether a speaker's conversational participation entitles them to influence a conversation, antecedent oppression may entitle speakers to moves other speakers are not entitled to make. An account of when this happens goes far beyond the scope of this article. However, we can formulate a general framework characterizing when an interlocutor might legitimately reject a speaker's attempt to influence conversational direction.

Ishani Maitra points out four possible reasons one might legitimately deprive someone of language's benefits: doing so might be necessary to achieve a greater good, securing the benefit is too costly, the person was never entitled to the benefit, and the person consented to the deprivation (Maitra Reference Maitra2009, 332). Maitra argues that communicative success is rarely legitimately overridden. However, speakers’ entitlements regarding communicative success and perlocutionary influence differ. The latter are much more easily overridden for valid reasons, legitimizing one's refusal to follow a speaker's attempt to guide the conversation. Maitra's conditions provide a way to examine legitimate reasons for preventing perlocutionary influence.

A speaker lacking entitlement or consenting to deprivation legitimizes one's refusal to follow the speaker's attempt to guide the conversation. A speaker might never have been entitled to influence conversational direction. Here, simple examples include courtroom witnesses. Even expert witnesses cannot change a lawyer's line of questioning. So even if an expert witness believes lawyers neglected relevant information, the witness cannot shift the conversation to that topic. The norms governing courtroom testimony do not entitle witnesses to influence the conversation as they wish. Conversational participation alone does not always generate a defeasible entitlement to influence conversational direction; the nature of the conversation and the participant's role in the conversation matter also. This most often happens in formal settings, such as meetings where a participant's expertise determines what they can contribute. Often, participation requires consenting to these constraints. However, one may find oneself in a conversation where strict norms govern conversational moves without consent, for example, as a defendant in court.

Regarding greater good and costliness of securing an entitlement, we can expand Maitra's ideas to conversational goal and overriding values. Certainly, greater good and costliness can override a speaker's entitlement to influence the conversation, but I propose two conditions more relevant to conversations that render an interlocutor's refusal legitimate: the attempted move conflicts with the conversational goal, and an overriding value grounds the refusal.

Conversations have goals. In the examples above, those conversational goals were pedagogical, playful, and constructive. A conversation's goal determines appropriate moves within that conversation. A conversation's goal can change as the conversation runs its course. However, given the conversation's goal at a specific time, one cannot expect abrupt change to the conversation without a good reason. The Frankenstein example reveals how conversational goals influence whether the teacher's refusal to follow the student's recommendation is legitimate. Suppose the teacher's goal was to cover specific content during the allotted time, and the teacher rightly believed discussing Frankenstein's motivation would take too much time. Here, the student still has a defeasible entitlement to change the conversational topic. However, changing the topic would prevent the class from completing the material in the allotted time. So, the change would have conflicted with the conversational goal and overridden the student's defeasible entitlement. If we change the conversational goal again, we get a different result. If the goal were for the students to better understand Frankenstein, getting through the content in the allotted time might not override the student's entitlement. Yet if most of the class could not follow the discussion because it was beyond the class's ability, the teacher could legitimately refuse to follow the student's recommendation. The conversational goal plays an important role here. An interlocutor can legitimately refuse to follow a speaker's attempt to guide a conversation when the attempted move conflicts with the conversational goal.

Relatedly, some values override active conversational entitlements. For example, time can be such a value. An interlocutor can refuse to follow a speaker's attempt to change the conversation when participants’ time constraints prohibit following the attempt, such as a long question in the final minutes of a lecture. Other values might include a topic's appropriateness given the participants. For instance, many topics are not appropriate when children are present. Similarly, introducing a sensitive topic after trauma may not be appropriate. So an interlocutor may refuse to discuss a topic in order to respect other participants. In each example, allowing the speaker to change the conversation would be too costly given the overriding values, so the interlocutor legitimately refuses the move because doing so achieves a greater good.

Although these values may override a participant's defeasible entitlement to influence the conversation, they do not always override those entitlements. Ultimately, the values overriding speakers’ entitlements are those generating a greater good than the conversational change itself. Objective criteria exist here. Participants should not misleadingly appeal to values, for example, politeness, simply to avoid an uncomfortable conversation. Instead, we question whether preventing the speaker from changing the topic brings about a greater good than allowing it. The answer to that question partly determines whether the interlocutor's refusal to follow the attempted conversational change is legitimate. Interlocutors will, however, make mistakes.

When interlocutors make mistakes, those mistakes do not justify refusing to follow the speaker's attempted conversational shift. Manifestly, assumptions based on social identity often produce such mistakes. Here, the interlocutor makes an assumption about the speaker based on an identity prejudice rather than evidence about the speaker. The interlocutor has not evaluated the speaker or whether the speaker should be followed; the interlocutor employs an unfounded assumption. This differs from the hearer refusing to follow the same speaker because experience has revealed ample evidence that the speaker cannot competently discuss the topic. So the interlocutor must have defeasible evidence that the speaker's attempt to direct the conversation is inappropriate given the context. The identity-prejudice issue is compelling because when an interlocutor has those entrenched beliefs, nothing the speaker does re-establishes her ability to influence the conversation, at least not in the direction the speaker intends. Moreover, entrenched identity prejudice makes the interlocutor resistant to new evidence. So the speaker cannot change her demoted standing within the conversation, at least on the current topic. Thus, perlocutionary silencing illegitimately deprives speakers of perlocutionary influence on a conversation to which their participation defeasibly entitles them.

In summary, conversational participation generates norms entitling speakers to perform perlocutionary acts influencing the conversation. These entitlements come from conversational participation rather than content. A speaker's defeasible entitlement to influence the conversation can, however, be overridden by an interlocutor legitimately refusing to follow the speaker's attempted move. An interlocutor legitimately refuses to follow a speaker when the speaker is not entitled to make that move, when the speaker consented to giving up that entitlement, when the attempted move conflicts with the conversational goal, or when an overriding value grounds the refusal.

Black Lives Matter, Women, and Perlocutionary Silencing

The current account of perlocutionary silencing identifies a linguistic, speech-related harm at the perlocutionary level. When speakers experience perlocutionary silencing, they are illegitimately denied access to conversational moves to which their linguistic agency and conversational participation entitle them. Although this analysis may not classify some examples as silencing that others might so classify, the analysis captures a serious harm, including two important examples of perlocutionary silencing: #AllLivesMatter and #NotAllMen.

The Black Lives Matter movement (hereafter BLM) campaigns against the systematic oppression of Black people. Speakers assert that Black lives matter, intending further perlocutionary effects, including eliminating systematic violence against Black lives. BLM also seeks discursive change, generating a conversation concerning the underappreciated value of Black lives. Speakers using “All Lives Matter” or “#AllLivesMatter” respond to BLM.Footnote ¹⁰ Speakers using “#AllLivesMatter” rely on the truth that all lives matter but make a discursive move blocking BLM's goal. Uses of “#AllLivesMatter” thwart BLM's discursive move by insinuating BLM ignores the value of all lives, privileging one group. Thereby, uses of “#AllLivesMatter” derail the conversation. So, when speakers use “#AllLivesMatter,” they silence BLM by illegitimately frustrating its discursive perlocutionary goal.Footnote ¹¹ Yet uses of “#AllLivesMatter” do not silence all perlocutionary goals. For example, BLM intends to change hearers’ behavior as it relates to ending Black oppression as a perlocutionary goal. Arguably, uses of “#AllLivesMatter” delay or prevent realizing this goal. This goal, however, is not a discursive goal. Thus, although uses of “#AllLivesMatter” cause perlocutionary frustration regarding changing behavior that would end oppression, it does not silence BLM regarding that perlocutionary goal. Perlocutionary silencing requires a discursive perlocutionary goal.

Like “#AllLivesMatter,” speakers introduced “#NotAllMen” as a derailment tactic, but this time against women highlighting pervasive abuse. When women highlight violence by men, they foreground needed change, a perlocutionary goal. Some men respond by drawing attention away from that conversation toward their own fears regarding false accusations. These speakers respond using “#NotAllMen” because not all men abuse women. Thus, “#NotAllMen” works as a defense tactic. Men who never abused women act defensively instead of compassionately. Consequently, those saying “Not All Men” derail the conversation.Footnote ¹² When women reveal assault, they highlight its prevalence. When men respond by emphasizing not all men abuse women, they shift the topic from women's experiences to how some men feel blamed for others’ actions. Here, again, speakers attempt a particular discursive move: Women use “#MeToo” among other locutions to identify the abuse they have experienced; these are illocutionary acts. But by using “#MeToo” to highlight the harms women endure, women also seek a conversational change focusing on the systematic violence they experience, a perlocutionary goal. Speakers using “#NotAllMen” illegitimately prevent that perlocutionary conversational move. So, as uses of “#AllLivesMatter” perlocutionarily silence BLM, uses of “#NotAllMen” perlocutionarily silence women. Both Black Lives Matter and women point out systematic violence and attempt to make discursive moves turning the public conversation to oppression, but speakers using “#AllLivesMatter” and “#NotAllMen” stop those conversational moves.

Further reflection on the BLM example offers greater insight into relevant perlocutionary failures. One might question whether BLM constitutes the right perlocutionary goal for perlocutionary silencing. BLM might rely on a conversational implicature, and if perlocutionary frustration involves a failed conversational implicature, it relies on participants sharing appropriate background assumptions. Thus, uses of “#AllLivesMatter” might not qualify as perlocutionary silencing because BLM's perlocutionary failures constitute nonlinguistic failures.

Successful conversational implicatures require conversational participants to share relevant assumptions as background knowledge. When conversational participants possess different background knowledge, hearers are less likely to interpret speakers correctly. For instance, if we assume most people believe that Black lives matter, the hearer might wonder why someone would say Black lives matter. This uncertainty generates the problem. The speaker hopes the hearer recognizes they intend more than merely asserting Black lives matter: they indicate the underappreciated value of Black lives. But other interpretations could explain the speaker's utterance. The hearer might think the speaker means only Black lives matter. Here, the hearer assumes the speaker would have said all lives matter if they thought all lives mattered. Since they did not, they must mean only Black lives matter. The speaker never intended this. But since the speaker relies on a conversational implicature, they cannot expect to influence the conversation based solely on their participation and the entitlements it brings. Instead, they base their reasons on beliefs about the interlocutor. So, the objection would go, BLM reaches their intended perlocutionary goal only when the implicature, an intermediate perlocutionary goal, first succeeds. Since conversational participation does not guarantee a successful conversational implicature, neither does it guarantee the conversational change from the implicature.

BLM's claims do not occur within a vacuum. Assumptions lurk in the background. Specifically, a conversation already exists regarding police shooting unarmed Black men and women. Speakers intend to advance this conversation. So, mere conversational participation entitles the speaker to turn the conversation to the value of Black lives, though these assumptions alone cannot guarantee the speaker will successfully direct the conversation.

Given these assumptions and the charge regarding how uses of “#AllLivesMatter” silence speakers using “Black Lives Matter,” we can now respond to the objection. Since police violence and the death of unarmed Black men and women remain salient, BLM protesters have defeasible reasons to believe they can turn the conversation to the value of Black lives. Others respond by using “#AllLivesMatter” as counter-speech. Instead of examining BLM's claims, asserting all lives matter blocks conversational moves, derailing BLM's intended conversational shift. Specifically, speakers asserting all lives matter reframe the discussion from injustice to the BLM movement's legitimacy. This change requires BLM to defend itself against charges of being dangerous, racist, or antipolice. Thus, speakers using “#AllLivesMatter” block further dialogue regarding injustices against Black lives. They force a specific interpretation of “Black Lives Matter” and set the background for further conversation. Ultimately, using “#AllLivesMatter” thereby distracts from what BLM protesters intended, like the recalcitrant questioner. “#AllLivesMatter” not only introduces a new conversational topic but illegitimately shuts down BLM's desired conversation. If a hearer misunderstands BLM's intended implicature, the hearer should more appropriately ask why the speaker feels she must make such a claim rather than derail the conversation. The hearer could open further discussion by questioning the speaker. Instead, the hearer does not try to understand the speaker. Thus, the hearer illegitimately dismisses the speaker without fully recognizing her agency. This occurs partly because the social system downplays racism and stymies openly discussing race-based issues. So structural racism enables speakers using “#AllLivesMatter” to derail the conversation and determine legitimate moves. These uses of “#AllLivesMatter” serve as counter-speech rather than cooperatively engaging with BLM. So uses of “#AllLivesMatter” qualify as silencing because uses of “#AllLivesMatter” derail the conversation, preventing BLM's discursive move, a discursive move that BLM speakers are defeasibly entitled to based on conversational participation.

Objections

One might wonder what my analysis adds to our understanding of silencing. We can interpret this worry in two ways. First, this analysis reduces perlocutionary silencing to phenomena that others identify. Second, perlocutionary silencing, here identified, collapses the distinction between locutionary, illocutionary, and perlocutionary silencing because it characterizes them as the same. Both objections deserve consideration.

In some respects, my analysis resembles views others defend. Jose Medina and Gaile Pohlhaus Jr. defend similar views when discussing epistemic injustice (Medina Reference Medina2012; Pohlhaus Reference Jr and Gaile2014). Both Medina and Pohlhaus argue that when an interlocutor engages a speaker in an epistemic exchange, the interlocutor grants the speaker some degree of epistemic agency. Epistemic agency entitles the speaker to norms governing the epistemic exchange. For Pohlhaus, access to these norms entitles speakers to perform certain speech acts, and for Medina, access entitles speakers to undertake sophisticated epistemic roles, for example, questioning, formulating hypotheses, and assessing and interpreting opinions. So speakers should be able to perform relevant speech acts and take on relevant epistemic roles if no legitimate overriding reason exists. Their positions, though, seemingly utilize the same features the current account does: Conversational participation entitles speakers to make certain conversational moves, performing speech acts and taking on roles. When interlocutors prevent those moves, speakers are harmed. My account of perlocutionary silencing identifies the same harm. So one might worry that no difference exists between Medina's and Pohlhaus's accounts and my account.

Yet my account differs in its level of generality. Medina and Pohlhaus focus on epistemic exchanges. Epistemic exchanges generate epistemic norms. So the interlocutor harms the speaker as an epistemic agent because the speaker's epistemic agency entitles them to certain epistemic roles. In Medina's and Pohlhaus's examples, interlocutors disrespect speakers’ epistemic agency, causing epistemic harms. Perlocutionary silencing, however, extends beyond epistemic harm. Interlocutors can illegitimately prevent speakers from achieving perlocutionary goals without a credibility deficit. Both Medina and Pohlhaus identify ways interlocutors illegitimately prevent speakers from directing a conversation within epistemic exchanges. However, conversational participation grants speakers access to norms governing the conversation more generally because conversational exchanges more generally generate norms. All conversations generate norms enabling phenomena similar to what Medina and Pohlhaus identify: instances of a more general pattern. My account elucidates that general pattern. Whereas Medina and Pohlhaus emphasize how epistemic exchanges generate conversational norms and harm speakers as epistemic agents, my account reveals how similar harms arise more generally, harming speakers as linguistic agents, not just as epistemic agents. Silencing prevents speakers from doing something their linguistic agency and conversational participation defeasibly guarantee. So, although similar to Medina's and Pohlhaus's accounts, my account offers a more general picture of the specific harm they identify.

By distinguishing between linguistic agency and epistemic agency, we can move beyond responding to this objection and further clarify perlocutionary silencing's harm as uniquely harming speakers as linguistic agents.Footnote ¹³ Linguistic agency is essentially the capacity to act using language, exercising one's agency using speech. As noted above, linguistic communication allows for further abilities and interests. Linguistic communication is necessary for and uniquely achieves nonepistemic stakes essential for human flourishing, for example, friendship, social intercourse, and expressing desires and refusals.Footnote ¹⁴ Since linguistic communication and conversational participation uniquely achieve these interests and are necessary for human flourishing, conversational participation and being an active part of a conversation matter. For this reason, we need an account of the harm that speakers experience that does not collapse into the narrower epistemic approach Medina and Pohlhaus take.

Separating epistemic capacities from linguistic capacities proves difficult because linguistic communication requires putting linguistic knowledge to use. Indeed, almost every action requires using knowledge. However, we need an account of illegitimate linguistic thwarting because, as noted above, we can achieve some important human interests only through linguistic communication. Surely, epistemic stakes pertain here, but those epistemic stakes differ from basic social interaction and social need. Consequently, linguistic harms result. Despite the epistemic dimensions in many situations, we can identify situations where interlocutors illegitimately block speakers’ perlocutionary moves resulting in harms not, or not primarily, harming the speaker as an epistemic agent. Prominent examples of epistemic harms include a hearer's identity prejudice resulting in attributing a credibility deficit to the speaker. Consequently, the speaker cannot fully utilize their epistemic agency. Conversational participants assess each other's credibility and use those assessments to determine access to conversational moves. However, even when a hearer does not attribute a credibility deficit to a speaker but still illegitimately blocks the speaker's conversational move, the hearer may still primarily harm the speaker as an epistemic agent.Footnote ¹⁵ The hearer may not thwart the move for epistemic reasons but might still harm the speaker as an epistemic agent if the speaker determines to put her knowledge to work for her. This is not simply about testimonial injustice. Something else could happen here. If the speaker knew something important and felt compelled to share that with her interlocutor but was denied that move because the speaker lacked nonepistemic social capital, we might still say the interlocutor harms the speaker's epistemic agency. I want to say a linguistic harm occurs here, and silencing and epistemic injustice often overlap in their occurrences. When a speaker is silenced, and it prevents the speaker from using their epistemic agency, an epistemic harm and a linguistic harm both occur. Yet to illustrate that these harms are distinct, an example will help, especially one where the epistemic harm either does not occur or, at least, is not the primary harm.

Suppose a group of friends get together reminiscing about college. As they tell stories they strengthen their friendship. When these friends get together, they tell the same stories and their conversations take the same form: someone mentions a funny event, the conversation moves to that story, different friends share their memories about that event, then someone else brings up another funny story, and they move on to that one. Ultimately, their conversational goal is strengthening their friendship by reminiscing with nostalgia. However, as each friend moves the conversation by bringing up good times they shared, one friend tries to turn the conversation to a time where the friends had a dinner party by saying, “Remember the time Karen stabbed Sean in the hand with a fondue fork during our fondue feast?” Another friend responds by saying “Oh yeah, that was good, but remember the time Darrin stole Aaron's car?” Here, conversational participation entitles the first friend to change the conversation to the memory he wanted to discuss. But before he could shift the conversation, the conversation had already moved to another topic because someone else dominated the discussion by introducing another memory. Here, the harm may seem minor, but if this continues throughout the conversation, it reflects a general disregard for the speaker as a conversational participant, that is, the interlocutor disrespects the speaker's linguistic agency. The speaker has nothing epistemically important to share. The speaker does not attempt to show he remembers a better story. The speaker merely wants to reminisce with friends, to which the conversation and topic entitle him.

Someone might object here, claiming the interlocutor harms the speaker as a friend rather than merely as a conversational participant. However, the speaker's entitlement to influence the conversation does not come from the friendship but from his conversational participation. The interlocutor's behavior is problematic for a friend because the interlocutor harms the speaker as a linguistic agent, whether as a slight in a one-off instance or more seriously if part of a general pattern.

When interlocutors illegitimately block speakers, they may harm them as epistemic agents, but they always harm them as participants in the linguistic exchange. As participants in exchanges, speakers have defeasible entitlements generated not because they are epistemic agents, but because they are speakers participating in conversations.Footnote ¹⁶

The second objection raises the worry that the current account collapses the distinction between locutionary, illocutionary, and perlocutionary silencing. The current analysis of perlocutionary silencing seems very similar to an account of illocutionary silencing. Essentially, the harms appear the same, collapsing the distinction between illocutionary and perlocutionary silencing. So why distinguish between them? Why not prefer an account of harm that does not rely on the distinction between illocution and perlocution but can explain both phenomena without having to make the distinction? This worry can be seen more clearly by comparing the current account with an alternative. Quill Kukla (writing as Rebecca Kukla) argues that when one's social identity systematically prevents them from producing their intended speech act, they suffer discursive injustice. More specifically, discursive injustice arises when a speaker, using appropriate words and tone and possessing the relevant entitlements, intends to perform a speech act, but the hearer fails to recognize the speaker's intention to perform that speech act (Kukla Reference Kukla2014, 445). On Kukla's account, the social differences a speech act makes determine its performative force (443). How hearers interpret or respond to an utterance determines whether the speaker succeeds in executing their intended speech act. Consequently, no agent completely controls her utterances’ performative force. Discursive injustice unjustly distorts an utterer's performative effects, preventing speakers from performing the speech acts they are entitled to produce. Discursive injustice, however, seemingly captures perlocutionary silencing without appealing to perlocution. This is a problem because discursive injustice provides a simpler account of linguistic harms that speakers experience.

Perlocutionary silencing appears to describe a discursive injustice that Kukla has already identified. Discursive injustice captures silencing without distinguishing between locutionary, illocutionary, and perlocutionary acts. Since perlocutionary silencing wrongs speakers similar to how locutionary and illocutionary silencing harms speakers and can be captured under “discursive injustice,” we have a distinction without a difference, better classified under the unitary discursive injustice. So, unless identifying perlocutionary silencing helps better understand the wrong, we should embrace the more general discursive injustice, which unites silencing with further nonsilencing harms to speakers’ linguistic agency.

Despite discursive injustice uniting certain linguistic harms, a meaningful distinction between perlocutionary and illocutionary silencings exists. The Frankenstein example reveals the distinction. There, the teacher treats the student's speech act as a recommendation. So the speech act succeeds despite the teacher treating the student unjustly. Kukla might respond by saying the teacher never took the recommendation seriously, so it lacked its intended performative force. However, this cannot be right. In the example, the recommendation succeeded qua recommendation but unsuccessfully changed the discourse as recommended. Discursive changes occurred compatible with a recommendation. If successful speech required perlocutionary success, ubiquitous discursive injustice would occur. Even accepting Kukla's account of discursive change constituting successful speech acts, we ought not to ignore perlocution. A meaningful distinction remains. The recalcitrant questioner reveals we can discursively respond to an intended assertion as an assertion, but further discursive injustice occurs. The speaker's assertion succeeds, but the interlocutor prevents further dialectical (perlocutionary) moves with their assertion. Kukla might think they identify the harm the recalcitrant questioner causes because the questioner discursively blocks the speaker, causing a discursive injustice. Yet the assertion succeeded, so the degree of linguistic agency the interlocutor recognizes differs from illocutionary silencing.

Since the distinction reveals differing degrees to which hearers recognize and ignore speakers’ agency, the distinction matters because it supplies explanatory power. The linguistic agency that speech acts require can be restricted or violated differently. Different types of speech acts reflect different intentions that speakers have: speakers intend to utter certain words with certain meanings, speakers intend to perform certain speech acts, and speakers intend their speech acts to have certain effects on their audiences.Footnote ¹⁷ So, although the general harm remains the same across locutionary, illocutionary, and perlocutionary acts, namely illegitimately restricted linguistic agency, the details differ. At the locutionary level, a speaker can expect to utter sentences without intimidation or physical hinderance. At the illocutionary level, a speaker speaking clearly, intending their words to have their standard meaning, to an attentive and competent listener should expect their utterance will be understood. At the perlocutionary level, far more background conditions and assumptions go into determining whether the speaker can expect success. Silencing simply prevents speakers from using their linguistic agency. All instances of silencing illegitimately restrict the speaker's agency, harming them, but each step from the locutionary act requires additional background conditions for expected success. Despite discursive injustice providing a unified analysis, we should draw the more nuanced distinction.

Perlocutionary acts require significantly more background knowledge and assumptions than do locutionary or illocutionary acts when determining whether we should expect their success. When speakers perform perlocutionary acts, they use their linguistic agency. When an interlocutor illegitimately interrupts a speaker's perlocutionary intention, the interlocutor disrespects the speaker's linguistic agency. Yet that disrespect differs from an interlocutor illegitimately interrupting a speaker's locutionary or illocutionary act. By distinguishing between locutionary, illocutionary, and perlocutionary silencing, we capture an important difference, and we better identify silencing's harms.