1. INTRODUCTION
Some injustices occur because of language. Speakers use language to exclude, offend, or embarrass audiences. In extreme cases, speakers use language to silence: they exploit aspects of language use that prevents others from fully succeeding in performing their intended speech-acts. Rae Langton (Reference Langton1993) and Jennifer Hornsby (Reference Hornsby1995) made many philosophers aware how pornography silences women's speech-acts by arguing that Austin's speech-act theory explains such silencing. For example, violent pornography that eroticizes refusal can influence expectations in sexual contexts. These expectations might cause a man to interpret refusal as consent. When a man no longer recognizes a woman's attempt to refuse his sexual advances because he is under the influence of such pornography, the woman is silenced. To explain this, Langton argues that when something outside the speaker's control prevents hearers from recognizing her intended speech-act, the speech-act fails, silencing the speaker. Specifically, the woman's speech-act fails because the hearer did not recognize her communicative intention. Thus, pornography silences women by stymieing their ability to make clear their intentions. Despite such a plausible explanation of silencing, language's silencing power requires further analysis.
Silencing sometimes occurs even while hearers recognize speakers' intended speech-acts, and Langton and Hornsby's accounts cannot explain these cases. Specifically, their accounts cannot explain some instances of silencing arising from what Miranda Fricker calls testimonial injustice. Testimonial injustice occurs when a speaker receives a deflated level of credibility from her audience because of prejudice on the hearer's part. In extreme cases, testimonial injustice silences speakers' assertions while the hearers recognize the speakers' intended speech-acts. So, instead of a man not recognizing a woman's intention to refuse his sexual advances because of the influence of violent pornography, a man might have such little regard for the credibility of women in sexual contexts that although he understands the woman's intentions, he does not take her utterance seriously. Thus silencing her refusal. In short, testimonial injustice disables speakers even when their hearers recognize their intentions. By identifying this extreme testimonial injustice, Fricker draws our attention to a different form of silencing. Since the hearer recognizes the speaker's intention, Langton and Hornsby's accounts cannot handle such silencing. Moreover, if different forms of silencing require different analyses, they will likely require different responses. So, if we want to rectify unjust silencing, we must understand the underlying linguistic phenomenon and how it allows us to respond to silencing.
I propose that the dialectical account of assertion, wherein assertion is characterized by its normative effects, namely those arising from the speaker's commitment to advocate for the asserted proposition, best explains silencing where hearers recognize speakers' intended speech-acts and provides a way to begin rectifying this silencing.Footnote 1 In so doing, I show that we can fill the lacuna by applying theories of assertion to silencing. The argument proceeds by first demonstrating that the dialectical account can explain this form of silencing. I then consider three further accounts of assertion, assertion characterized by the speaker's attitude, assertion as a proposal to the common ground, and assertion defined by its constitutive norm(s), and how they do not explain this form of silencing. I proceed by first laying out silencing as it differs from what Langton considers. I then explain silencing in terms of the dialectical account. With this positive account of silencing in place, I go on to show the three other accounts of assertion fail to explain silencing arising from testimonial injustice. I conclude by showing how the dialectical account provides a start for rectifying silencing from testimonial injustice.
2. SILENCING
2.1 Silencing Uptake
Communication requires coordination. For communicative success, speakers and hearers must coordinate their thoughts. Occasionally coordination fails, silencing speakers. Catharine MacKinnon (Reference MacKinnon1987, Reference MacKinnon1993) famously argued that pornography silences and subordinates women. Following MacKinnon's argument, Rae Langton (Reference Langton1993) explains how this happens. Violent pornography depicts rape and eroticizes refusal. Since this pornography portrays women as enjoying unwanted and forced sexual encounters, male viewers may think this is normal. Thus, “Someone learning the rules of the game from this kind of pornography might not even recognize an attempted refusal” (324). Men influenced by such pornography no longer recognize refusal as a legitimate move in the “sex game.” In this context a woman can only consent, not refuse. Langton further points to final year schoolboys (high school seniors) unaware they did anything wrong, when in reality they raped girls. These girls were silenced because their speech-act of saying ‘no’ was disabled. They were not interpreted as refusals. Disabling a woman's ability to refuse takes away her autonomy and authority. By not recognizing a woman's attempted refusal as such, hearers silence her. Despite a woman saying “No!” a man might continue his sexual advances, raping her. Langton argues that in this case the woman's attempted refusal fails, silencing her speech-act. When some external force prevents a speaker from coordinating her thoughts with her hearer, her speech-act is silenced.
To understand silencing, we must understand the conditions necessary for a successful speech-act. J.L. Austin (Reference Austin1962) called these “felicity conditions.” When a speaker utters a sentence, her speech act can fail in a variety of ways. For example, I might enter a cafe and attempt to order a cappuccino by uttering ‘cappuccino.’ One way my order might fail would be if the barista added too much milk to the espresso. I ordered a cappuccino, but got a latte. I successfully ordered a cappuccino, but my order was not fulfilled. In Austin's terms, although my illocutionary act succeeded, my perlocutionary goal failed. Alternatively, I might utter ‘cappuccino’ intending to order a cappuccino, but the barista mistakes my order for a question, responding with ‘one-third espresso, one-third steamed milk, and one-third wet foam.’ In this instance, there was a communicative failure. My illocutionary act was not fully successful. According to Austin, my illocutionary act failed because I could not secure uptake. Uptake is a felicity condition requiring that hearers recognize speakers' communicative intentions.
Langton (Reference Langton1993) and Jennifer Hornsby (Reference Hornsby1995), individually and together (Hornsby and Langton Reference Hornsby and Langton1998), argued that Austin's speech-act theory explains why the refusal example qualifies as silencing, arguing that uptake is necessary for explaining silencing. According to Hornsby and Langton, successful speech-acts require uptake, which “consists in the speaker being taken to be performing the very illocutionary act that, in being so taken, she (the speaker) is performing” (Hornsby and Langton Reference Hornsby and Langton1998: 25). Thus, communication requires receptiveness. Successful uptake does not entail that the hearer agrees with what the speaker says. Instead, uptake requires the hearer's grasping the communicative act the speaker intends to perform. When hearers meet this requirement, Hornsby and Langton claim the speaker's utterance succeeds. Thus they say, “A speaker tries to do an illocutionary thing; a hearer recognizing that the speaker is trying to do that thing is then sufficient for the speaker to actually do this” (my emphasis, 25).Footnote 2 Indeed, uptake is also essential for successful illocutionary acts. Consider a warning. If the hearer misidentifies the speaker's intention, the warning fails, i.e. the speaker has not warned the hearer. This can happen because the hearer thought the speaker was joking. Since uptake requires that the hearer recognize the speaker's intention to warn, when uptake fails, the warning fails (26).Footnote 3 Likewise, uptake failure explains how pornography silences women. When uptake fails, attempts at illocutionary acts manifest illocutionary disablement – the speaker fails to satisfy the conditions for the intended speech-act (26–7). In our example where a woman refuses in a sexual context, uptake fails because of an external force – the influence of violent pornography. Since the man misidentifies what she tries to do with her words, her utterance of ‘No’ fails despite her intention to refuse the man (27). When felicity conditions go unsatisfied, speech-acts misfire (27). As a felicity condition, when uptake fails, the speaker is silenced.Footnote 4
By characterizing silencing as illocutionary disablement, rather than the speaker's failure to reach her perlocutionary goals, this account of silencing reveals silencing's harm. When a hearer understands, but ignores, a speaker's refusal, the hearer harms the speaker by forcing her to do something she does not want to do. Yet, if the speaker is prevented from effectively communicating her refusal, i.e. illocutionary disablement, there is a further harm. As Maura Tumulty points out, characterizing silencing as illocutionary disablement is compelling because it explains how women are harmed beyond the direct harm of assault when their refusals are ignored (Tumulty Reference Tumulty, Crasnow and Superson2012: 222). These women are deprived of authority and autonomy to express their desires. More generally, analyzing silencing as illocutionary disablement reveals how speakers are harmed as agents in a communicative exchange.
Despite the plausibility of their analysis of silencing's harm, by focusing on warning and refusal, Hornsby and Langton overlook silencing where uptake occurs, but the speech-act still fails. For warnings and refusals, illocutionary success makes uptake both necessary and sufficient, but other speech-acts require more than uptake for success. For example, a bet requires the bet taker accept the offer (Austin Reference Austin1962: 9). A successful bet requires more than the hearer recognizing the speaker's intention. So, uptake, although necessary, is not sufficient for betting. Now, not accepting a bet does not silence the bet maker by illocutionary failure, but we can consider further speech-acts requiring more than uptake, viz. assertion. Specifically, hearers can fully grasp the content and intention of the speaker attempting an assertion, but by not taking the speaker seriously the speech-act fails, silencing the speaker.
2.2 Silencing When Uptake is Satisfied
If assertion fails because hearers don't take speakers seriously, this failure may result from a credibility deficit attributed to the speakers. When hearers inappropriately attribute credibility deficits to speakers and do not take them seriously, Miranda Fricker (Reference Fricker2007) calls this testimonial injustice. Fricker's central case of testimonial injustice looks at credibility deficit based on identity prejudice. Hearers attribute a credibility deficit to speakers when they attribute lower credibility levels to speakers than their evidence recommends. This combines with identity prejudice when the hearer bases his prejudice on the speaker's social identity (28). She argues that prejudice can distort how hearers perceive speakers and consequently distort hearers' credibility judgment of the speaker (36). When this results in hearers giving speakers deflated credibility levels, testimonial injustice occurs (1). For instance, hearers believing women are intuitive and emotional in their decisions may attribute a credibility deficit to them because the hearer assumes intuition and emotion should not influence decisions. Such testimonial injustice excludes the subject from trustful conversation. Since the hearer fails to recognize the speaker as trustworthy, he excludes her assuming she lacks the capacity as a giver of knowledge (44, 53).Footnote 5 So, when hearers attribute credibility deficits based on identity prejudice, hearers, although understanding the speaker fully, might fail to take the speaker seriously.
Despite the harm in testimonial injustice so characterized, Fricker points out that it takes on a more extreme form: silencing. Accordingly, testimonial injustice can cause “fundamental communicative dysfunction” (139). Here, a credibility deficit causes speakers' utterances not to register with hearers' testimonial sensibilities. Thus, hearers never hear the speaker (140). So, hearers silence speakers through testimonial injustice by attributing them an unfair credibility deficit. Following Fricker, we can use Langton's refusal example to make this point. Imagine a man attributing a drastic credibility deficit to women because he views them as irrational. He thinks women only make decisions based on emotions. Since the man attributes such a drastic credibility deficit to women, when one speaks, her words never register with his testimonial sensibilities. Although she utters meaningful words, the man does not respect her as an epistemic agent, so he never takes her words as giving him a reason for believing her (140–1). So, if this woman is not fully successful in refusing the man's sexual advances because he attributes her a credibility deficit, we get a different explanation of silencing from Langton. In Langton's example, the woman's credibility never comes into question; her refusal fails because uptake fails. In Fricker's example, the woman attempts to refuse the sexual advances, but the man's negative identity prejudice causes him not to recognize her utterance as a meaningful contribution to the conversation. Since the man does not take the woman's utterance as worth considering, he silences her refusal. Despite focusing on refusal, similar examples arise for assertion.
Fricker takes this silencing example as extreme testimonial injustice, but groups affected by negative identity prejudices are often silenced. Women regularly experience silencing. For instance, women commonly utter sentences with no response, while in the same conversation a man receives credit for the same idea. When a credibility deficit causes this, these women are silenced. Assuming the conversational participants do not take the woman seriously because they attribute a credibility deficit from a negative identity prejudice, her attempted assertion fails. She attempted to inform or convince her interlocutors, but they did not recognize her contribution to the conversation. So not only did the woman fail to convince her audience, but it was as if she never said anything at all. Such an injustice requires an explanation.
Before proceeding to explain silencing assertion, we should first consider silencing's nature and what qualifies as a successful explanation of silencing. Silencing takes different forms depending on the conversational dysfunction. We want to understand silencing speech-acts. In particular, those speech-acts speakers intend to perform in uttering linguistic expressions by their conventional force, often called the illocutionary act. Some classify this silencing as illocutionary-disablement – for some reason the speaker's intended illocutionary act fails. This alone does not qualify as silencing, for refusing a bet from a child need not qualify as silencing nor harmful, so we need a narrower category of speech-act.Footnote 6 Instead we can follow Ishani Maitra's development of Hornsby's communicative acts (Maitra Reference Maitra2009: 321). Hornsby classifies communicative acts as those speech-acts where success depends on hearers recognizing the speaker's intention to perform the act. Warnings qualify as communicative acts because a speaker warns a hearer only if the hearer recognizes the speaker is trying to warn her. With this in mind Maitra argues for silencing as communicative disablement. That is the speaker cannot fully successfully perform her intended communicative act because her audience failed to recognize her informative intention or recognized the informative intention but interpreted it differently (327–8). However, not recognizing the speaker's intentions does not characterize silencing that interests us. Successful assertion requires hearers to recognize the speaker's intentions to some degree, so assertion qualifies as a communicative act. So, a satisfactory account of silencing assertion must explain how a speaker does not successfully perform her assertion while the hearer recognizes her intentions. Moreover, Austin pointed out that when we utter sentences, we perform different acts. The intended effect a speaker has is the perlocutionary act. In our example, hearers fail to believe the speaker, frustrating the perlocutionary goal. When a speaker fails to convince her audience because the latter provides a counterexample, this frustrates her perlocutionary goal, but does not harm her. Our example differs significantly because the perlocutionary goal fails precisely because the communicative act fails. The speaker cannot make her assertion because of a conversational dysfunction. The conversational dysfunction disables her communicative act. This form of silencing occurs at the illocutionary-communicative level, but not because the hearer fails to recognize the speaker's intention. Here, silencing assertion differs importantly from other silencings. Although assertion qualifies as a communicative act and requires recognizing speakers' intentions for success, such success is not sufficient for a successful assertion.
Any acceptable account of assertion must explain silencing where the hearer recognizes the speaker's intentions. Likewise, a satisfactory explanation cannot merely explain this silencing by pointing out the speaker's intended effect failed. So, a hearer disbelieving the speaker does not automatically silence her; the hearer may reject the assertion for good reasons. So, a satisfactory explanation must explain how a speaker's intended assertion is not fully successful, qua assertion, where the hearers recognize the speaker's assertoric intentions and the failure does not depend on the hearers not believing the proposition the speaker intends to assert. Because testimonial injustice involves testimony and testimony involves assertion, we want to know how theories of assertion explain this form of silencing. The dialectical account of assertion based on speakers' commitments provides the most promising explanation of silencing.
3. ACCOUNTING FOR SILENCING
3.1 Dialectical Account of Assertion
The dialectical account of assertion characterizes assertion by its normative effect in relation to other speech-acts (Brandom Reference Brandom1994; Rescorla Reference Rescorla2009; MacFarlane Reference MacFarlane, Brown and Cappelen2011). In so doing, an utterance's normative status within a language game determines its status as a speech-act. Specifically, when a speaker asserts a proposition, she undertakes and acknowledges doxastic commitments carrying normative social status for which the question of entitlements arise (Brandom Reference Brandom1994: 142). The normative status changes by the speaker taking on responsibility for the content of her utterance. Specifically, the speaker must justify the assertion, or retract it if none can be given, when questioned or challenged. Assuming the speaker can justify her assertion, she is entitled to take the content of the assertion for granted as the conversation progresses. So, when hearers regard speakers as authoritative and do not challenge them, this entitles both speaker and hearer to the proposition asserted and any proposition logically entailed by it as their conversation progresses. This is the entitlement. The speaker also becomes responsible for her asserted proposition. In this way, she must provide reasons for the proposition if challenged. As Robert Brandom points out, we specify assertion by the significance of speakers undertaking the commitment to justify their asserted propositions (188). By making an assertion, a speaker alters the conversation's normative status licensing others to re-assert the claim because the speaker takes on justificatory responsibility (Brandom Reference Brandom1983: 642). Accordingly, assertion changes the normative status and requirements upon the speaker differing from other speech-acts such as promising. For assertion, the speaker commits to advocating for and defending the proposition asserted if challenged, i.e. the speaker can be held responsible for the assertion's content. While for a promise, the normative commitments include fulfilling the promise. Correspondingly, we characterize assertion as a move in the game of giving and asking for reasons. When speakers commit to advocate for a proposition and take on that responsibility, they make their speech-acts assertions.
However, this normative status of commitment within the language game arises only because those attributing or acknowledging such commitments take the utterance as so (Brandom Reference Brandom1994: 161). Accordingly, assertion is like notarization. Only within the notarizing practice does a notary's stamp carry its status. Likewise, only within the practice of giving and asking for reasons does a speech-act qualify as an assertion. More importantly, assertion alters the normative status only because conversational participants attribute commitments or entitlements (166). The conversational participants' actual attitudes determine the act's normative status (184). Thus participants allow speech-acts to qualify as assertions. In this way Brandom says, “we accord [the status of assertion] to the utterances of children, madmen, and foreigners only to the degree we take them to understand the inferential relations within which the responsibility and authority of asserting arises” (Brandom Reference Brandom1983: 644). Communal response makes an utterance an assertion on the dialectical account, since communal response determines the conversation's normative status, namely to what the participants are entitled. Commitments relating to assertion only arise in the context of reasoned discourse. As Brandom points out, we take on commitments when and only when those we grant the authority to hold us responsible take us as authoritative. For example, suppose a child believing himself responsible for a baby claims responsibility for the baby. If someone else bears responsibility for the child and the baby, the child lacks the responsibility no matter what he does, because those who can grant him the responsibility do not. So, just as a babysitter is responsible for a child only if the primary caregivers make that person responsible, a speaker is responsible for her assertion only if those she grants the status of interlocutor, i.e. having the authority to hold her responsible, take her as authoritative.Footnote 7
3.2 Silencing
Since the dialectical account characterizes assertion as altering the normative features relating to commitment and entitlement, when they do not change, no assertion occurs. Thus, assertion is part of a conversational exchange in the same way pitching is a move in a baseball game. When a participant makes a move within the game, the game's status changes – the other players recognize the move and act accordingly. But sometimes an event does not change the game's status. For example, despite the strike zone being objectively specified, the umpire decides whether a pitch is a strike or a ball. Regardless of the ball's trajectory, the umpire's call, which entitles players to, or sanctions players from, making further moves, determines its status. Similarly, since asserters grant their hearers the authority for holding them responsible for their assertions, speakers give their hearers an umpire-like status or authority.Footnote 8 So, if those present for a speech-act fail to recognize, either intentionally or unintentionally, the normative change typically resulting from that utterance, the speech-act does not occur. By ignoring, or refusing to accept, the change in normative status following an assertoric utterance, conversational participants silence the speaker rendering the attempted assertion null and void.
We can directly apply this to Fricker's account of testimonial injustice and explain how victims of testimonial injustice are silenced by it. On her picture, there are social climates where some members of the community lack credibility because of negative identity prejudices, and for certain subject matters their words fail altogether to register their hearers' testimonial sensibility (Fricker Reference Fricker2007: 141). When testimonial injustice takes this form, neither what speech-act the speaker attempts to perform nor its content is lost on the hearer. The hearer fully grasps the speaker's intention to assert, but the hearer gives no credence to what the speaker says. As a result, the conversation's normative status does not change. The hearer does not credit the speaker with committing to a content because he assumes the speaker lacks the authority to justify that content if questioned. Thus when someone accepts a black man's testimony only after a white corroborator comes forward, the black man is silenced because although his utterance was likely understood, the normative status of the context did not change by his utterance. From this we see the dialectical account provides a framework that explains and realizes the form of silence Fricker presents. So the dialectical account provides an account of assertion that explains silencing from testimonial injustice.
4. INADEQUATE ALTERNATIVES
Despite the dialectical account's success explaining silencing arising from testimonial injustice, three other well-known accounts of assertion stand out as possible co-rivals. John MacFarlane (Reference MacFarlane, Brown and Cappelen2011) divides theories of assertion into four general categories. These categories classify assertion based on the speaker's attitude or intentions, effects on common ground, assertion's constitutive norm, or commitment associated with assertion. Following MacFarlane, we call the first three the attitudinal account, the common ground account, and the constitutive norm account respectively. The last we have called the dialectical account. A complete discussion of each is not possible here. Instead, I shall provide each account a brief explanation sufficient for approaching silencing.
4.1 Attitudinal Account
The attitudinal account characterizes successful assertoric communication by speakers' intended perlocutionary effects. Similar to how Grice analyzed speaker meaning as speakers intending to generate an effect in their hearers by their utterance, speakers elicit an attitude in their hearers (Grice Reference Grice1989: 220). Speakers' assertions succeed when their hearers recognize what they were trying to do with their utterances. The attitudinal account characterizes assertion by speakers' intentions to make their hearers aware of their intentions. Bach and Harnish (Reference Bach and Harnish1979) most recognizably proposed such an account. They characterize assertion as follows: In uttering a sentence, a speaker asserts a proposition if the speaker intends for the hearer to take the utterance as a reason to think the speaker believes the proposition and intends for the hearer to believe that proposition (15, 42). Thus, an assertion succeeds when the speaker's intended effects are met and the speaker intends for her assertion to affect the hearer's beliefs. Thus a speaker makes an assertion that P by uttering a sentence when the speaker (1) intends for the hearer to believe, based on the utterance, that she, the speaker, believes that P, and (2) intends for the hearer to believe that P. So, making an assertion depends on the hearer recognizing that the speaker intends the hearer believe the proposition the utterance was meant as a reason to believe.
4.1.1 Objection
Since the attitudinal account of assertion characterizes assertion as the expression of an attitude, the attitudinal account seems best situated to explain silencing as Hornsby and Langton consider it. For on the attitudinal account, successful assertion depends on the audience recognizing the speaker's intention to assert. This success, however, is the uptake condition. When we apply uptake to silencing currently under consideration, such a condition is not sufficient. Silencing here characterized takes for granted that an utterance occurs and the audience recognizes the speaker's intended speech-act, but a credibility deficit causes the speech-act to fail. Accordingly, illocutionary disablement occurs. But the attitudinal account cannot capture this because as long as the hearer recognizes the speaker's intention to assert, the speech-act succeeds, i.e. silencing does not occur.
In response, one might object that if someone does not believe that she can cause others to believe what she says, then she can't intend for them to believe it. If she cannot form the intention, then we might explain silencing on the attitudinal account. However, this only applies where the speaker can't form the intention. If speakers typically expect their hearers not to ignore them, this response fails. When speakers attempt to make assertions, hearers typically recognize their intentions to do so. Where speakers know their audience will not accept their assertions and have no reason to think they will recognize the intention, most will not even try to make an assertion. If a speaker still performs an utterance where she expects hearers will not recognize her intention, she is either irrational, or trying to do something besides make an assertion, such as avoiding sanction when an utterance is expected.
4.2 Common Ground Account
The common ground account of assertion characterizes assertion by its effects. Robert Stalnaker (Reference Stalnaker1978) analyzes assertion by its effects on context, what he calls the common ground. The common ground is the set of propositions that the conversational participants accept. When speakers make assertions, they attempt to change the context in two ways. First, and more basically, assertion alters the common ground to accommodate that the speaker uttered the words she did. Typically, any conspicuous change in the conversational surroundings change the common ground in a similar way. So, simply uttering a sentence changes the common ground. The second change, however, is unique to assertion and central to the common ground account. If hearers don't object to the statement, they add the asserted proposition to the common ground (Stalnaker Reference Stalnaker1978: 86). So speakers make assertions by proposing propositions to the common ground. If accepted, participants incorporate the proposition into the body of accepted information among the conversational participants (Stalnaker Reference Stalnaker1998: 99). But successful assertion does not require hearers accept the proposition into the common ground. To assert a proposition is to propose that it be added to the common ground. Participants can challenge the assertion, and if successful, the asserted proposition is not added to the common ground. Successful challenges do not cancel assertions.
4.2.1 Objections
We might interpret Stalnaker in two ways regarding silencing. First, we might interpret him as allowing that a hearer silences a speaker by not accepting her proposition into the common ground. Yet, when a speaker proposes a proposition to the common ground, other participants can reject the proposition, e.g. they have direct evidence the proposition is false. This does not qualify as silencing specified above, for not believing the proposition alone does not qualify as silencing. Second, we might interpret him as allowing that a hearer silences a speaker by preventing the speaker from proposing a proposition to the common ground. Accordingly, speakers make assertions by making a proposal to the common ground and add to the common ground that a proposal was made. Speakers make proposals when participants update the common ground reflecting that a proposal was made. Initially, this seems plausible. If the hearers didn't know the speaker made a proposal, then they cannot update the common ground. This would prevent the first change to the common ground, and what Sarah Murray (Reference Murray2014) calls direct updating. This, however, leaves out the case under consideration. The correct account must explain silencing where uptake occurs, but the hearer does not take it seriously. That is, the hearer recognizes the speaker's attempt to make an assertion, but by not giving the speaker credit the hearer does not update the context – common ground – in the way she would if the assertion were taken seriously. Thus, this approach requires committing to the implausible position that a hearer could recognize that the speaker made a proposal, yet not update the common ground accordingly. Such an individual would be almost irrational: He recognizes the speaker proposed something but chooses not to believe that it was a proposal. For on the common ground account, the context updates based on the participants' shared beliefs. So, if the hearer recognizes the utterance as a proposal, he would believe a proposal was made.Footnote 9
4.3 Constitutive Norm Account
An alternative account of assertion, the constitutive norm account, characterizes assertion by how one properly asserts. Since speakers deserve praise or criticism for making assertions, norms must govern assertion's proper performance. Such norms are constitutive norms. An act's constitutive norms are essential to that act, i.e. necessarily the rule governs every performance of the act (Williamson Reference Williamson2000: 239). We define assertion then as a move in the language game by the rules governing its proper execution. Thus, we articulate the speech-act's constitutive norms. Timothy Williamson, however, points out that constitutive rules do not provide necessary conditions for the speech-act (240). Instead, constitutive rules govern how to properly carry out an act: one does not fail to assert by violating a constitutive norm of assertion, but one asserts incorrectly by violating the norm(s). The norm of assertion, then, individuates assertion by explicating proper assertion, and the rule uniquely individuates the speech-act. So, assertion is a self-contained practice governed by its constitutive norm(s) – the norm(s) one must satisfy to participate properly in a practice. Such an approach characterizes the norm of assertion as NA:
NA: One must: assert p only if φ.
Where ‘ p’ is a proposition and ‘φ’ is some condition necessary for properly asserting p.
Theorists sometimes characterize this norm making ‘φ’ belief (Bach and Harnish Reference Bach and Harnish1979) or truth (Weiner Reference Weiner2005, Reference Weiner2007), but they often characterize it by some epistemic standard speakers must satisfy for proper assertion. Different standards have been proposed, for example knowledge (Williamson Reference Williamson1996; DeRose Reference DeRose2002), certainty (Stanley Reference Stanley2008), justified belief (Kvanvig Reference Kvanvig, Greenough and Pritchard2009), rational belief (Douven Reference Douven2006), or reasonability for believing one has knowledge (Neta Reference Neta2009). More recently, Sanford Goldberg (Reference Goldberg2015) has argued for a robustly epistemic context sensitive norm. For the present purpose, we need not take any one as best characterizing the constitutive norm account of assertion. Instead, we can follow Goldberg's initial formulation without committing to any specific standard other than it being epistemic. So, we take a schematic approach stipulating that whatever the norm of assertion is, it must be robustly epistemic, i.e. any speaker satisfying the norm has enough epistemic authority to warrant her assertion. Thus, the Robust Constraint:
The Robust Constraint: S must: assert p only if S satisfies epistemic condition E with respect to the proposition p. (207)
Where, ‘E’ replaces a description of the relevant epistemic standard, and this standard must be robustly epistemic. Hence, the Robust Constraint characterizes the robustly epistemic norm of assertion that purportedly warrants assertion.
On the constitutive norm approach to assertion, we define assertion by the rules (norms) governing its proper execution. So as Michael Rescorla points out, this approach characterizes assertion apart from other speech-acts: assertion is a self-contained practice governed by a robustly epistemic rule for proper, i.e. warranted, assertion (Reference Rescorla2009: 102). Consequently, we characterize assertion in isolation without reference to any other speech-act. Despite sharing “language-game” terminology and normative considerations, the constitutive norm and dialectical accounts significantly differ and the dialectical account contrasts directly with this very approach.
To start, the constitutive norm account characterizes assertion by how to properly assert, or, more precisely, what to assert, while on the dialectical account, we can only make sense of assertion by the normative change it creates, namely how the speaker must respond to challenges or questions. On the constitutive norm account, norms constrain what one should assert, e.g. only assert what you know. So, speakers assert by performing a speech-act governed by the norm(s) of assertion, i.e. the rule(s) governing all and only assertions (Goldberg Reference Goldberg2015: 11). In contrast, the dialectical account characterizes assertion by normative standing or the normative change assertions create. So, on the dialectical account constitutive norms arise regarding how to respond when challenged, e.g. defend when legitimately challenged or retract (Rescorla Reference Rescorla2009: 103). The dialectical account allows constitutive norms for assertion, but they are not necessary and don't do the explanatory work characterizing assertion. According to the dialectical account, the commitments a speaker takes on by performing her speech-act makes the speech-act an assertion (Rescorla Reference Rescorla2009: 105). In this way, the dialectical account makes assertion part of the larger practice of reasoned discourse, and accordingly, assertion can only be fully characterized by its normative relations to other speech-acts.
To further illustrate the difference between the dialectical and constitutive norm accounts, consider someone that appears to make assertions but responds by weakening the claim. Such a person appears to make assertions, but if challenged responds with ‘It's possible.’ Whatever he's doing is significantly weaker than assertion. Yet, the constitutive norm account classifies this as an improper assertion because the speech-act was governed, i.e. judged, by some epistemic norm. In contrast, the dialectical account does not provide an automatic response. Accordingly, it may not be an assertion because, when challenged, the speaker does not respond by justifying the claim with the justification an assertion requires, yet the speaker may think he's doing this but merely provides inadequate justification when challenged.
4.3.1 Objections
At first glance, the constitutive norm account looks like another plausible contender for explaining silencing. Since a robustly epistemic norm governs assertion, when a speaker fails to satisfy that norm, her assertion is flawed. But testimonial injustice occurs when hearers attribute the speaker a credibility deficit regardless of whether she fails to satisfy the Robust Constraint. Instead, the hearer unjustifiably assumes the speaker has not satisfied the Robust Constraint and judges that her assertion should not be taken seriously. Consider what Fricker has in mind. We assume a social climate where some community members lack credibility with the hearer because of a negative identity prejudice. When the victim of a negative identity prejudice attempts to make an assertion, the hearer perceives her as untrustworthy and does not believe her. Fricker says the speaker is silenced in extreme cases, but the constitutive norm account cannot explain this. Here, most plausibly the hearer attributes a credibility deficit to the speaker to the point that the hearer assumes the speaker violates the Robust Constraint. Consequently, the speaker's assertion would not qualify for transmitting knowledge.
So, the constitutive norm account explains why a credibility deficit causes hearers to judge assertions inadequate for transmitting knowledge, but this phenomenon is not silencing. Since the hearer judges the speech-act according to the robust constraint, the norm governs the speech-act making it an assertion. The constitutive norm account limits the explanation to an assumed inappropriate assertion. Consequently, instead of explaining silencing, we have an explanation of how the hearer concludes that the speaker made an illicit move in the conversation. So, when we attempt to apply the constitutive norm account to silencing resulting from testimonial injustice, it cannot account for silencing as it interests us.
4.4 The Robust Constraint's Return
Even though the constitutive norm account fails to explain silencing, it deserves further consideration. Recently, Goldberg (Reference Goldberg2015) argued that the constitutive norm account explains assertion's central features, including ethical features. For he claims the Robust Constraint explains testimonial injustice and the wrong in not taking others' assertions seriously. To do so, he relies on the Robust Constraint being mutually manifest to language users. By this he means participants in a speech exchange are aware that assertion has a robustly epistemic norm (96). The Robust Constraint's mutual manifestness entails that when a speaker makes an assertion her audience is entitled to assume that the speaker has the relevant epistemic goods to satisfy the Robust Constraint. So, assertion entitles the audience to form expectations that go hand-in-hand with speakers satisfying the Robust Constraint. Likewise, when a speaker makes an assertion where it is mutually manifest that success depends on cooperation and coordination, this entitles the speaker to expect that the hearer, as a co-participant, will inform her when rejecting her assertion (181). For such cooperation and coordination require conversational participants be open regarding their responses to co-participants' utterances.
Since speakers form these expectations, the hearer disrespects the speaker when he disregards the assertion. In such situations, Goldberg claims the hearer does not give the speaker her due and offends against the speaker's implicit claim at having the relevant epistemic authority (182). By not taking the speaker seriously, the hearer disrespects the speaker. Goldberg thinks a hearer takes an assertion seriously when reacting to the assertion by believing its content or taking the assertion as an opportunity for believing its content (182). If hearers do not take assertions as an opportunity for believing their contents, they disrespect the speakers. For example, they might require more justification than they would require of speakers against whom they lacked an identity prejudice. By not granting assertions the status of conveying epistemic authority, the hearer does not grant the speech-act and speaker the statuses they deserve. When a hearer fails to take an assertion seriously, he morally harms the speaker by disrespecting her (185–6). However, testimonial injustice disables the speaker, not just disrespects her, and if the hearer disables the speaker, he has silenced her. This may, however, be a terminological difference, so we should clarify Goldberg's disrespect.
Since the robust constraint is mutually manifest, both speaker and hearer realize the speaker appeals to her epistemic authority by making an assertion. In particular, Goldberg says,
To ignore or dismiss an assertion on insufficient grounds, then is to disrespect the speaker as an epistemic subject by inappropriately repudiating her claim to relevant authoritativeness. (Goldberg Reference Goldberg2015: 188 original emphasis)
But this does not account for silencing. Goldberg recognizes other harms occur when one ignores or dismisses an assertion. For instance, the speaker is wronged as a participant in a joint epistemic effort, or activity (188). The harm arises as an insult because the speaker and hearer engage in a two-way exchange where the speaker requests a response on behalf of the hearer. This harm arises because participants form cooperative expectations. When a hearer rejects an assertion on insufficient grounds, a uniquely epistemic wrong occurs. The speaker and hearer do not just participate in a general cooperative exchange; they participate in a joint epistemic activity. This activity requires that participants recognize each other as epistemic agents, i.e. givers of knowledge based on their authority. By rejecting the speaker's assertion without considering it, the hearer rejects the speaker as a participant in the exchange but also rejects the speaker's claim to authority. Because of the Robust Constraint, when a speaker makes an assertion, she makes two requests. First the speaker requests that the hearer consider the proposition asserted. But, second, the speaker also requests that the hearer consider her as an epistemic agent, i.e. someone claiming to have the epistemic authority to warrant the assertion. When someone rejects assertions based on an identity prejudice, he disregards the speaker as an agent. Without question this harms the speaker.
Now, is Goldberg's disrespect testimonial injustice, and does it silence the speaker? Goldberg needs hearers' actions, or inactions, to deprive the speaker of an active voice in public discourse. To some extent the hearer excludes the speaker from the deliberative process, but the Robust Constraint's mutual manifestness does not obviously explain this or make such exclusion silencing. Goldberg can certainly claim that the Robust Constraint explains epistemic disrespect, and thus testimonial injustice. But nothing in the account is necessary for explaining how the speaker loses her active voice in the deliberative process. Thus the Robust Constraint itself is otiose in explaining silencing.
5. RECTIFYING SOME TESTIMONIAL INJUSTICES
Having seen that only the dialectical account explains why extreme testimonial injustice can cause silencing, we should consider whether it tells us anything about remedying silencing. To do this, let's begin with a worry: the dialectical account gives too much power to hearers over speakers. By not updating the conversational score, hearers silence speakers, preventing their assertions. Although the dialectical account entails that hearers can do this, we should not interpret this as a problem. Rather, this is a strength. We might hope the ideal account of assertion deprives hearers of the power to silence speakers, but such a theory ignores real-world injustices. Since we have presupposed that silencing prevents speakers' assertions, the best account of assertion should accommodate this presupposition. Relatedly, one might worry that the dialectical account ignores the individualistic component of assertion: although assertion is a public practice, individuals do it. The dialectical account, however, explains both individual and public components. On Brandom's formulation, assertions depend on what group comprises the conversational participants. Linguistic communities have subcommunities. As Brandom points out, sometimes we distinguish subcommunities by the authority they acknowledge. Although a subcommunity shares and recognizes certain entitlements, other subcommunities may not recognize those same entitlements (Reference Brandom1994: 242). As a result, members of one subcommunity may not be able to challenge the assertions of members in other subcommunities, and may not be able to make intersubcommunity assertions. Consequently, when conversational contexts include participants from different subcommunities and some attribute responsibility for the utterance and others do not, the utterance's status differs relative to the subcommunities. When this happens because systematic identity prejudice within one subcommunity leads them to judge the other as lacking credibility, the speaker both asserts and is silenced – the speaker makes an assertion relative to the subcommunity of which she is a member, but is silenced relative to the community of which she is not a member. This does not rectify silencing, but provides us with a more nuanced picture as overlapping subcommunities provide the answer for rectifying silencing.
The dialectical account does not merely account for the silencing that accompanies testimonial injustice; it also provides a means to rectifying it in some contexts. When none of the hearers take the speaker's attempted assertion seriously, the dialectical account entails that the intended assertion does not occur because the required deontic change does not occur. In some contexts the speaker can do nothing to remedy this. When a woman attempts to make an assertion in a meeting and her boss does not register the assertion, in some sense no one else registers an assertion, i.e. they do not update their deontic scorecards. We can compare conversational authority figures to referees in sports. The crowd waits for the referee to determine the status of the play. In a business meeting, employees might wait for their boss to decide the status of utterances: if the boss does not recognize the utterance as worthy of consideration, no one considers it. This happens partly because the boss holds authority in that context, not unlike referees in sports. But when participants recognize that the speaker's attempted assertion was dismissed illegitimately, they can seek to update the deontic score by making an assertion that defers credit to the original speaker. The speaker rectifies the injustice because the update has two parts: (1) the speaker's assertion provides conversational participants with entitlements and responsibilities for the assertion, but (2) by saying “As so-and-so said a moment ago” the speaker passes the responsibility on to the original speaker.
When testimonial injustice silences a speaker, she fails to perform her intended speech-act. The dialectical account provides sufficient tools to rectify this because it characterizes assertion by changes within the conversational score. Each participant tracks relevant changes as the conversation progresses. When a speaker makes an assertion, hearers update the conversational score, but not merely by adding the asserted proposition to the common ground. According to the dialectical account, when a speaker makes an accepted assertion, she is entitled to use the proposition asserted and anything entailed by that proposition as premises as the conversation continues. So assertion entitles conversational participants to propositions extending beyond the asserted proposition. This is similar to David Lewis's (Reference Lewis1979) accommodation for presupposition. When a speaker makes an assertion requiring a presupposition, conversational participants update the conversational score accommodating the presupposition. But since the dialectical account makes assertion depend on responsibility and entitlements, when a conversational participant makes another participant responsible for a proposition because of her original utterance, the original utterance acquires the status of assertion. So a speaker who utters ‘As Olive said a moment ago …’ adds to the conversational score both the proposition that Olive said and that she said it. In a context where hearers attribute speakers a credibility deficit, this is a way for other conversational participants to make sure those speakers retain their voice. Since the conversational score updates as it does, the responsibility and entitlement is credited to Olive, through the current speaker. In this way the speaker makes an assertion based on testimony previously unrecognized.
This provides a further reason for accepting the dialectical account. Since the constitutive norm, attitudinal, and common ground accounts fail to recognize silencing as we've been discussing it, they do not provide a means for rectifying this particular harm. Nor can they explain why deferring benefits the original speaker. We should note this means of rectifying silencing does not work in every context, nor does it rectify testimonial injustice more generally. Fricker argues that by developing the virtue of epistemic justice we avoid testimonial injustice. This is fine advice. But we cannot force others to develop virtue. So in the meantime we shall do the best we can and rectify what wrongs we can. The dialectical account of assertion analyzes assertion in such a way that it explains silencing resulting from testimonial injustice, but also provides the tools for those that recognize this silencing to rectify it when we can. Even if hybrid formulations of the constitutive norm, attitudinal, and common ground accounts of assertion can account for silencing resulting from testimonial injustice, if they do not provide a straightforward means for rectifying silencing, we should prefer any account that primarily analyzes assertion as the dialectical account does, by normative change through the discursive process.
6. CONCLUSION
Testimonial injustice arises in many contexts. In extreme cases, testimonial injustice can silence speaker's assertions. Yet, such silencing is unique in that the communicative act is not fully successful despite hearers recognizing the speakers' intentions. Making sense of this form of silencing requires correctly understanding assertion, for the correct account of assertion must be consistent with this phenomenon. Having looked at four different accounts of assertion, only the dialectical account of assertion can explain silencing resulting from testimonial injustice. For this reason, the dialectical account of assertion is the correct view of assertion. As we saw, the attitudinal account cannot explain silencing arising from testimonial injustice because it makes recognition of the speaker's intentions sufficient for successful assertion, but the form of silencing we are considering silences the speaker even when hearers recognize her intentions. Likewise, the common ground view fails to capture silencing because as long as the speaker succeeds in making a proposal, assertion occurs. Finally, the constitutive norm account fails to capture silencing because as long as the Robust Constraint governs an utterance, it qualifies as an assertion, and when hearers attribute a credibility deficit, they judge the speaker by a robustly epistemic norm. However, the dialectical account explains silencing resulting from testimonial injustice, because a credibility deficit impedes the speaker's utterance from changing the normative status of the conversation. Moreover, the dialectical account provides a framework within which some speakers can rectify the silencing of others by deferring credit to speakers that are silenced.Footnote 10
