Epistemic Injustice in Political Discourses? The Problematic Concept of Authority in Langton’s Account of Pornography

. Through her silencing thesis, Langton has contributed to the study of epistemic injustice by highlighting a possible cause of such a phenomenon: She asserts that the pornographic representation of (straight) sexual relationships affects the felicity conditions of speech uttered by women, so this speech is not understood as an illocution by men. This fact arguably undermines women’s credibility, since their testimony is not even registered in men’s testimonial sensibility. However, this thesis entails problematic consequences from at least two standpoints. From a theoretical perspective, it enacts a circularity when it comes to the empirical individuation of the subordinative effects of pornography. I will point out that this problem arises from Langton’s substantive conception of power, i.e. from her notion of authority as an attribute which can be ascribed to preexisting subjects. From a political perspective, such conception of power allows Langton to performatively rank women as credible when testifying sexual violence, but it also leads her to silencing alternative political strategies, e.g. the ones proposed by Butler. Hence, I propose to consider this form of silencing as a specific kind of epistemic injustice, one that neutralises the performative value of political discourses.

exist… The idea that there is either located at -or emanating from-a given point something which is a power seems to me to be based on a misguided analysis" (1980, p. 198). Power should not be reified as an attribute that can be ascribed to preexisting subjects, as a privilege that stays at their disposal and constitutes hierarchical relationships within society. On the contrary, it should be conceived in an anti-voluntarist way, that is as a cultural and linguistic conditioning that operates through practices which, in turn, shape the very subjectivities of the individuals that unknowingly reproduce them through their usual behaviours. 3 In applying this notion of power to my analysis of Langton's essay "Speech Acts and Unspeakable Acts" (Langton, 2009b), I will focus on a specific way in which power conditions her argumentation. By drawing on William Connolly (1995) and Bonnie Honig's works (1991), I will assume that philosophical texts are never reducible to mere logical deductions, but can be also interpreted as performatives which naturalise or renegotiate 4 common presumptions about social reality. These presumptions, in turn, occasion political consequences that do not always conform to the author's intentions, but often prove to have conservative effects for what regards the elaboration of political strategies to contrast subordination. In Langton's case, the main social-ontological presumption is of course her substantive conception of power, and the possible detrimental effect it can exercise consists in presenting power as constituting rigid social hierarchies. This ultimately entails that disempowered groups are aprioristically supposed not to be able to autonomously claim their own rights.
On the basis of these presuppositions, I will analyse the aforementioned circularity not as a logical incoherence, but as a performative, which has the political function to denounce the subordination perpetrated by pornography. From this perspective, some aspects that are generally overlooked in the logical analysis of a text play an important role. In particular, as I will argue in the second section, figurative language allows Langton to present ambiguous examples of subordination as unambiguously instantiating illocutionary disablement. In this way, Langton reconstructs empirical cases of silencing to conceal the circularity enacted by her argumentation and to naturalise the idea that women's perspective on gender-related subordination is the most reliable description thereof. Langton's performative can thus be characterised as an illocution that ranks women as epistemically authoritative when testifying sexual violence. As a result, her contribution to the study of epistemic injustice does not only depend on her philosophical elaboration, but also on her political attempt to renegotiate the common conception of women's subordination.
This illocution also entails an unintended side-effect, since it implies that women's credibility depends on their lacking practical authority. As I will argue in the third section, this assumption leads Langton to aprioristically discrediting political proposals that aim to enable women to actively engage in the political debate. As an example of this tendency, I will quote Langton's response to Butler's criticisms and argue that the content of Excitable Speech is interpreted in constative terms, thus being deprived of its political potential. I conclude the article by describing this silencing act as an illocution which constitutes a specific form of epistemic injustice. As a matter of fact, unlike more common instances of such a phenomenon, it does not depend on a substantive conception of power, and does not entail an unbalance of authority between Langton and Butler. If anything, it highlights the reproduction of power relationships which are external to the immediate interaction between speaker and hearer.

Langton on Pornography: A Performative to Silence and Subordinate
Langton's silencing argument (2009) describes pornography as performing a speech act that she calls illocutionary disablement. However, in order to comprehend her thesis, it is first necessary to explain the meaning of the Austinian concept of illocution.
This notion is notoriously distinguished from locution (which refers to the mere faculty to utter words provided with referential meaning) since it qualifies linguistic agency (Austin, 1962). It thus involves the capacity to produce effects by means of language, but, in this sense, it has to be differentiated from perlocution: As a matter of fact, in this latter case effects constitute all the contingent (i.e. unpredictable and generally unintended) consequences that an utterance occasions in its audience, such as emotional responses, psychological associations, and so on. Effects are thus produced by perlocutionary acts in a causal fashion, in that they are not generally recognised as typical responses to such utterances. This latter feature is in fact what most characterises illocutions: unlike instances of perlocution, they are provided with a conventional meaning which is assumed as familiar to all the interlocutors and is thus associated with the expectation of some specific responses. So, when a speaker performs an illocutionary act, she presumes that everybody knows which reactions are intelligible as responses to that act, therefore she also exercises a normative conditioning on the way her audience will act after her illocution. Convention grounds the intersubjective character 3 An efficacious synthesis of Foucault's notion of power is provided by Butler. She presents it as a "shift from the subject of power to a set of practices in which power is actualized in its effects… Among the very practices that Foucault counters to that of the subject are those that seek to account for the formation of the subject itself" (1997, p. 79). In order to explain this point, she quotes Foucault's proposal to "grasp subjection in its material instance as a constitution of subjects" (Butler, 1997, p. 79). 4 By renegotiation I mean the performative act of addressing the audience's political imagination and moral convictions in order to change their expectations about what is legitimate or plausible when it comes to political intervention. The object of such expectations is what I label social ontology. of linguistic interactions, so that, for instance, when a bride says "I do" during a wedding ceremony, she performs the act of marrying: in saying those words, she immediately brings about the circumstance that her hearers recognise her as a wife, rather than, say, as a single woman (Langton, 2009b). 5 What kind of speech act does pornography perform? According to Langton, it is an illocution that can be determined in a twofold way: pornography is "first, verdictive speech that ranks women as sex objects, and, second, exercitive speech that legitimates sexual violence" (2009b, p. 40). In other words, pornography ascribes a socially significant identity to the class of individuals it labels, and by the same token it authorises the behaviour that is most appropriate to such an identity. As a result, since it qualifies women as sex objects, it normalises sexual violence, thereby conjuring it as morally unproblematic.
However, pornography does not simply change its hearers' beliefs about sexual practices -since this would be a mere perlocutionary effect. Rather, pornography silences women through the very act of presenting them as objects. Of course, this does not mean that it prevents them from uttering words -physical coercion is perlocution as well-but that it deprives their words of their illocutionary force. In this sense, women's speech is not intended as an action by men, it does not affect their expectations about what response is most significant in that context, so that such hearers will act just like no one has ever spoken: "Someone learning the rules of the [sexual] game from this kind of pornography might not even recognize an attempted refusal" (2009b, p. 58). That is because the pornographic distortion of women's speech would turn a woman's "No" into an expression of consent. Silencing thus operates "not simply by depriving speech of its intended illocutionary force, but by replacing it with a force that is its antithesis" (2009b, p. 59). Women's utterances are displaced from the context that is necessary for their illocutionary success, insofar as their meaning is adapted to the expectations enacted by sexual objectification. This phenomenon is what Langton calls illocutionary disablement: But there is a… kind of silencing that happens when one speaks, one utters words, and fails not simply to achieve the effect one aims at, but fails to perform the very action one intends… although the appropriate words are uttered, with the appropriate intention, the speaker fails to perform the intended illocutionary act. (2009b, p. 48).
In order to get a better understanding of this notion, it is convenient to explicit that illocutions succeed only if they satisfy some contextual requirements that Austin and Langton call "felicity conditions." Thus, since illocutionary disablement is described as neutralising an utterance's illocutionary force, it must be possible to analyse it by enumerating all the felicity conditions it suppresses.
Most evidently, the quoted passage asserts that pornography neutralises the speaker's capacity to display intentions. The importance of this felicity condition can be deduced from the methodological role it plays in Langton's argumentation: she assumes the consequentiality between the speaker's intentions, the action performed and the realisation of the intended effects as a linguistic paradigm, and employs it to trace back illocutionary disablement from a failure to produce the intended effects. However, it is patent that such a failure may be a case of perlocutionary frustration as well, insofar as it can result from the fact that the hearer has grasped the speaker's intentions, but does not want to conform to them.
Hence, the absence of intended effects cannot work as an evidence of illocutionary disablement, unless it simultaneously highlights a misinterpretation of the speaker's intentions. This entails not only the suppression of another essential felicity condition of speech acts, which consists in securing uptake, but also the interruption of the intersubjective relationship of communication: indeed, whereas the speaker's intention to perform an illocution is a necessary, but not sufficient condition for its success, the possibility to secure uptake about this intention is what actually allows for the felicity of an illocution. After all, no hearer can interpret -or even misinterpret-an illocution without simultaneously ascribing a correspondent intention to its speaker. 6 Nevertheless, even in this case some specifications are needed, because misunderstandings can be caused by a great deal of factors. In particular, uptake is itself dependent on further contextual requirements (e.g. previous utterances, social conventions…). Hence, in order to demonstrate that a failed interpretation instantiates illocutionary disablement it has to be proven that this failure is conditioned by the lack of one last felicity condition, one that has a specifically political meaning, that is authority. In fact, Langton asserts that one's social position conditions the meaning of her utterances: the word "Fault!" pronounced by an umpire will change the scoreboard, but the same exclamation will be interpreted as the expression of a personal opinion if it is uttered by a bystander, because this lacks the appropriate authority to bring about the same effect as the umpire.
Similar examples about authority recur in Langton's argumentation, such as the way an utterance's meaning changes according to whether it is pronounced by a slave or by his master, by a legislator or by a citisen (Langton, 2009b, pp. 36-37). What is common to all these examples is that authority is conceived as an exclusive privilege of certain subjects, who are thereby distinguished from others according to a powerful/powerless binary. This follows from the fact that the very notion of power is adapted from institutional contexts and applied to informal linguistic 5 In this sense, Langton agrees with other scholars who contend that illocution is a linguistic means to create social facts (Lewis, 1979) and thus shapes the context that allows for the significance of future illocutions (Witek, 2013). See Langton, 2009b, pp. 52-53;2017, p. 151. 6 Langton, 2009b, p. 33: "Speech acts are a subset of actions in general, so there will always be some description under which a speech act is intentionally performed." On this point, Langton seems to assume that the possibility to render one's intention public is the main possibility condition of illocutionary success. For a more detailed account of this background requirement, see Hornsby, 1995. interactions, so the main difference between them is aprioristically obscured. Even though Langton explicitly admits that everyday linguistic interactions are not regulated by explicit rules, she does not conclude that they may depend on a different kind of power than the one that emerges from legally structured contexts. On the contrary, she substitutes law with convention (2009b, p. 53), thereby preserving the same radical differentiation between who has and who lacks authority. This binary is in fact one of the implications of what I formerly called substantive conception of authority, because it presents power as a discrete entity which can be identified as being or not being present in a specific subject. 7 This can also be deduced from the way authority is figuratively represented by Langton: it is conjured as a thing that can be destroyed ("[i]f pornography prevents her [a woman] from refusing" sex, it "destroys her authority," Langton, 2009b, p. 59); as a position which can be occupied by speakers (subordinating speech acts depend on "the condition that the speakers occupy a position of authority," Langton, 2009b, p. 42); as something that can be monopolised ("it may be that pornography has all the authority of a monopoly," Langton, 2009b, p. 46). All these descriptions tend to reify the idea of power, namely by presenting it as a quality that inheres to some subjects and allows for the intelligibility of their speech acts.
In this sense, power is not itself dependent on practices: it is not seen as a set of diverse, mobile and unstable relations which are reproduced by everyday linguistic interactions (Young, 2011;Foucault, 1980). It is rather the social practices which are rendered intelligible and efficacious by existing social hierarchies, so that these hierarchies are in turn conceived as transcendent in relation to the speakers' actual behaviours. This assumption eventually allows Langton to consider the problem of verifying the authoritative status of pornography as an empirical question: Since verdictives and exercitives are both authoritative illocutions, we know that the ability to perform them depends on the speaker's authority…. This means that in order to answer the question: "Does pornography subordinate?" one must first answer another: "Do its speakers have authority?". (2009b, p. 44).
Quite evidently, these questions depend on the hypothesis that pornographic speech is an instance of illocution, so it is also necessary to prove this latter claim in order to demonstrate that pornography's speakers have authority. So, after having evaluated different argumentative strategies, Langton proposes to adopt the following method: "[O]ne may argue that a speech act's effects are best explained by supposing that it has a certain illocutionary force" (2009b, p. 42). In other words, proving that the perlocutionary effects of pornography are those which would normally be expected from an illocutionary act of subordination is a necessary, but not sufficient condition to demonstrate that pornography is a subordinative illocution. There may still be different causes for such effects. For this reason, this hypothesis has to be confirmed by verifying that pornography's speakers have authority.
But what kind of effects can incontrovertibly evidence that pornography is an instance of illocution? Of course, the effects of illocutionary disablement. This in turn means that, in order to demonstrate "MacKinnon's claim […] that pornographic speech, in particular, silences the speech of women," it is necessary to show that producing pornography is a way "to use speech to disable speakers… to prevent them from satisfying the felicity conditions for some illocutions they might want to perform" (Langton, 2009b, p. 53). According to the reconstruction of the notion of illocutionary disablement I formerly provided in this section, this thesis implies that pornography makes it impossible for women to communicate intentions and secure the appropriate uptake for their speech acts.
In order to test the validity of the silencing argument, Langton applies her methodological assumptions to an empirical case. The protagonist of the episode she analyses is Linda Marchiano, who, after having played a central role in the pornographic movie Deep Throat, attempted to protest against the violence she experienced as an actress by writing a book called Ordeal. Nevertheless, her book was surprisingly exploited as a means of sexual entertainment. How can such a failure be explained?
The answer is simple. It is there [in a mail-order catalogue for adult reading] because it is pornography after all: here, in this context, for these intended hearers, the uptake secured is bound to be that of pornography…. She intends to protest. But her speech misfires. Something about who she is, something about the role she occupies, prevents her from satisfying protest's felicity conditions, at least here. (2009b, p. 55).
Langton highlights the fact that the perlocutionary effects of Marchiano's speech act contradict her intention to protest, and ascribes this failure to the fact that pornography altered the conventions regulating the discourse about sex. Pornography therefore distorts protest's meaning and impedes Marchiano to secure the appropriate uptake for her utterance, which arguably is one of protest's felicity conditions. Langton concludes her point by arguing that her example reflects a case of illocutionary disablement: just like an actor who fails to warn the public about a real fire during his performance, an ex-pornographic actress fails not just to be believed, but also to be properly understood in her denunciation of pornography's violence. (2009b, pp. 55-56).
However, the argumentation is ambiguous, because Langton limits herself to deducing the misinterpretation of protest from the fact that Marchiano's speech act causes sexual arousal. Sexual arousal is just a perlocutionary effect, and could be explained in many alternative ways than by reference to illocutionary disablement. Most evidently, Langton's analysis does not rule out the possibility that some readers may have recognised Ordeal's attempt to protest, but deliberately neglected it and decided to use it as pornographic material. One may even think that the very attempt to protest proved to be sexually attractive for some perverted readers. 8 From a methodological point of view, it is impossible to exclude these possibilities. It is also difficult to demonstrate that Marchiano failed to secure uptake for her speech act, if one does not previously demonstrate that her speech was interpreted as intending to perform a different action than the one she really wanted to perform. As a matter of fact, as I mentioned, it is impossible to (mis) understand a speech act without ascribing it an intention which reflects one's interpretation. But Langton provides no evidence of this misunderstanding.
Hence, the assertion that pornography sets up the rules in the language game of sex is unwarranted. So, what allows her to conclude that this alleged rule-setting performative "requires the premise that pornography is authoritative speech, otherwise it could not rank and legitimate" (2009b, pp. 59-60)? What does the term require mean? It cannot hint at a deduction of authority from pornography's illocution, because this presupposition has not been demonstrated. So, it must recall the assumption according to which "in order to answer the question: 'Does pornography subordinate?' one must first answer another: 'Do its speakers have authority? '" (2009b, p. 44). Authority thus proves to be the presupposition from which the illocutionary force of pornography is deduced.
This is actually what emerges from other works by Langton. It is worthwhile to consider her "Pornography's Divine Command? Response to Judith Butler" (2009c): here she refers to the empirical studies she quoted in "Speech Acts and Unspeakable Acts" as evidence that pornography alters its consumers' normative beliefs, and concludes that "[t]hese perlocutionary effects… are poorly explained by an assumption that pornography has no authority; they are better explained by an assumption that pornography does have authority" (2009c, pp. 110-111). However, in order to state that pornography's authority is the best assumption to explain the alteration it causes in its consumers' normative beliefs, it is first necessary to demonstrate that such an alteration is a linguistic phenomenon -more specifically, a perlocutionary effect of a previous authoritative illocution. This has not been stated by the studies that Langton quotes in her essay, which in fact refer to the evidence in question as a result of pornography's working as a cognitive stimulus, rather than as a speech act. 9 Hence, instead of assuming pornography's authority to explain a linguistic effect, Langton uses the same assumption to qualify an effect as linguistic, and just for this reason can she conclude that pornography has authority. Her argument is therefore circular.
I would like to argue that this circularity issues from the fact that Langton conceives authority in a substantive way. Indeed, only because she adapts this notion from legally structured contexts can she present power relationships as prior and independent in relation to linguistic practices. Her neglection of the differences between formal and informal contexts thus impedes her to formulate a sufficiently clear notion of authority for the latter ones. As a result, it is unclear how pornography's authority can be empirically individuated, because according to Langton's argumentation it can be just detected if one is able to individuate successful illocutions. Nevertheless, if the identification of an illocution depends on the fact that pornography requires authority, the reasoning begs the question. Given this difficulty, it is likewise unclear how authority conditions illocutionary efficacy. One may even ask whether it is authority that makes illocution felicitous, or whether it is felicitous illocutions that produce authority (Maitra, 2012). This may be the main reason why so many theorists doubt the possibility to empirically detect the phenomenon of illocutionary disablement described by Langton. 10 This problem can hardly be solved by providing a definition of power which is supposed to be stricter than Langton's one. For this reason, I will adopt Foucault's nominalistic approach to power by assuming that this cannot be described by abstracting from its contextual manifestation. From this perspective, power cannot be reified in inert social structures, nor can it be ascribed to preexisting subjects. On the contrary, it is immanent to a historically contingent set of social practices, that in turn shape the individuals' subjectivities (Foucault, 1980;Butler, 1997). This formative effect is what most differentiates this notion of power from the one I formerly called substantive: power is not something that stays at its holders' disposal and enables them to arbitrarily perpetrate violence toward disadvantaged groups. It is rather a normative conditioning, which operates through conventions regulating social interactions and which, by the same token, naturalises a social ontology which legitimises these conventions: it fixes the roles agents can play within everyday encounters, as well as the rights and duties associated with such roles; it shapes the context that allows for one's utterances' meaning, and thus circumscribes the cultural space that renders some experiences speakable, while rendering other unintelligible. 8 Some other possibilities are enumerated by Green (1998, p. 298) and by Maitra (2009, p. 314, note 10). 9 See Langton (2009b, p. 39). The plausibility of interpreting pornography as a cognitive stimulus is admitted by Langton in the essay "Scorekeeping in a Pornographic Language Game," which she wrote with Caroline West (Langton & West, 2009). However, she does never develop this hypothesis, because it is patent that, were it true, pornography would operate its effects without depending on social context, so that it could not be described as a speech act. 10 Among Langton's critics, see e.g. Green (1998) and Jacobson (1995). Among her supporters, see e.g. McGowan (2004) and Maitra (2009). Furthermore, Caroline West (2003) admits that Langton's theory has not been empirically confirmed.
It is therefore evident that this hypothesis allows for the possibility of there being social hierarchies; but it also conjures these hierarchies as strictly dependent on the actions performed by the subjects within them. Furthermore, these power relationships can hardly be represented according to the powerful/powerless binary, since the relational character of the practices that consolidate them often involves agents who are not directly identifiable as exercising power or being subjected to it (Young, 2011). 11 By assuming this notion of power, I do not intend to formulate an alternative account of pornographic silencing. Rather, I aim to demonstrate that Langton's account of pornographic subordination contributes to reproduce some gender-related presumptions, which may prove to have detrimental effects from a political point of view. To this purpose, I will analyse her essay "Speech Acts and Unspeakable Acts" again, but this time I will interpret it as a performative rather than as a scientific text.

Langton's Narrative Construction of the Pornographic Scene of Utterance
Recall that the circularity of Langton's argumentation was enacted by a specific proposition: the silencing phenomenon "requires the premise that pornography is authoritative speech" (Langton, 2009b, pp. 59-60). Indeed, this sentence hints at the possibility to determine a silencing act as a case of illocutionary disablement on the basis of its speaker's authority, even though this very authority has to be deduced from a previous act of illocutionary disablement. In this way, the whole argumentation excludes alternative explanations of the silencing phenomenon, because by considering the utterance about pornographic authority as a constative it proves to be incoherent. Nevertheless, this utterance can also be interpreted as a performative one, which has the specific function of denouncing pornography by persuading the reader that it has sufficient authority to legitimate sexual violence. In such case, neither the exclusion of alternative explanations for the silencing phenomenon nor the aforementioned circularity would invalidate the argumentation. As a matter of fact, this would have to be evaluated not by reference to its truth or coherence, but in virtue of its efficacy in producing its intended effects.
In supporting this thesis, I draw on the work of some postmodern philosophers who argue that political discourses are never reducible to pure logical deductions. On the contrary, they are necessarily substantiated by contestable presumptions that reflect a contingent social ontology -e.g. contents of collective memory or political imagination, moral convictions, and so on. However, on the one hand, such presumptions are necessary to justify a political thesis in terms of plausibility, realism, and the like; on the other hand, they are themselves unjustifiable (Connolly, 1995). Hence, political argumentations depend for their very intelligibility on a crucial ambiguity, because they are simultaneously constative logical deductions and performative attempts to renegotiate or naturalise commonly shared convictions 12 (Derrida, 2002;Honig, 1991).
In Langton's argumentation, the main social-ontological presumption can be identified in her substantive conception of authority, which in turn conditions her description of current social relationships. In particular, it grounds the following presumptions: first, women have the authority to decide on their own body and refuse sex; second, an unequal distribution of power unjustly disempowers them; third, social relationships must be adjusted in order to restore such a violated right. 13 As I will try to demonstrate, it is not clear whether the assumption that women (should) have authority is simply stated or rather instituted by Langton, so that her argumentation proves to coincide with the aforementioned performative ambiguity. In this section, I will analyse Langton's employment of figurative language. Thus, I will highlight that, in conjuring her narratives, she casts ambiguous examples of subordination, such as Marchiano's one, as unambiguously instantiating illocutionary disablement.
This expositive process allows Langton to present women's subordinated status as a vantage point on sexual violence, thereby persuading the reader that their testimony about this phenomenon is more credible than the one of other social groups -such as men and pornographers. I propose to consider this reappropriation of epistemic authority as a performative not just because it arises from representation rather than argumentation, but also because it is structured as an illocution that ranks women's testimony as credible by naturalising their standpoint on sexual violence. 14 An important role in this performative reappropriation of epistemic authority is played by the description of the conversation between pornography and its consumers. In particular, consider the following example: 11 Foucault illustrates this dynamic by saying: "[T]his power is not exercised simply as an obligation or a prohibition on those who 'do not have it'; it invests them, is transmitted by and through them; it exerts pressure upon them, just as they themselves, in their struggle against it, resist the grip it has on them" (1995, p. 27). The pressure in question refers to the conventions that regulate some specific practices, which include philosophical and political discourses (Foucault, 1981). Therefore, it is reasonable to expect that it can be traced also in Langton's works. 12 Of course, in asserting this, I allow that my argumentation is affected by the same performative ambiguity. 13 For an explicit illustration of the first of these three presumptions, see e.g. Langton (2009b, pp. 58-59): "A woman who prohibits sexual advances also has authority within the local domain of her own life, her own body… If pornography prevents her from refusing, then pornography destroys her authority as it twists her words." For the enunciation of the other two presumptions, see Langton (2009b, p. 63): "Women wish to be able to speak some important actions: to be able to refuse, to protest, or to give testimony. The speech of pornographers may prevent them from doing so. If it does, then it may be wrong for a government to allow pornographers to speak." 14 In this way, Langton adopts a partial perspective on sexual subordination, thereby making a choice she herself refused by saying that it would have compromised the impartiality of her method. See Langton, 2009b, pp. 43-4: "We might say that those women who take pornography to be subordination are in a better position to judge… But unless we privilege one group of hearers in this way, our result with this way of arguing will be inconclusive." The authors of pornographic speech are not mere bystanders to the game [of sex]; they are speakers whose verdict counts. Pornography tells its hearers what women are worth: it ranks women as things, as objects, as prey. Pornography tells its hearers which moves are appropriate and permissible: if it tells them that certain moves are appropriate because women want to be raped, it legitimates violence. (Langton, 2009b, p. 44).
Quite evidently, pornographic speech proceeds unilaterally and unambiguously from speakers to hearers. The duality of the interlocutors presents pornography and its consumers as playing complementary roles, respectively source and recipient of a predetermined performative meaning. In this way, Langton reenforces the idea that pornography straightforwardly produces its intended effects: as a matter of fact, it tells its hearers which moves are appropriate and permissible, and thus immediately legitimates violence. Formulations like this may be the actual reason why many critics misinterpreted Langton's thesis as asserting that pornography constitutes subordination without needing to satisfy any contextual condition (see e.g. Butler, 1997;Green, 1995;Vadas, 1987). This frequent misunderstanding does not depend on the explicit content of Langton's argumentation -quite the opposite, she often clarifies that the efficacy of pornographic speech depends on its hearers' social context. If anything, it highlights that this context is never described: its reconstruction is rather presented as an empirical problem, so the strictly philosophical analysis of pornography's functioning only requires to describe its effects on an abstract subject. This abstraction from social context conceals a difficult question: if pornography produces subordinative effects because of its operating within a context in which gender-related subordination is already institutionalised, what evidences that sexual violence results from pornography, rather than other practices?
This indeterminacy has a specific function in Langton's examples of pornographic speech. It allows her to present women as already incapable of resisting sexual subordination before the utterance of pornographic speech, even though pornographic speech is described as responsible for their subordination. Thus, women are always excluded from the conversation between men and pornographers. This is evident if one considers that, in describing the way pornography ranks women as things, Langton presents them as objectified not just as targets of sexual desire, but also as grammatical category: pornography speaks about women, while never speaking to them, i.e. without addressing them as speaking subjects. As a result, men come to conceive women as sexual instruments only because they are already unable to resist erotisation when pornographic speech is uttered.
This representational strategy is thus structured as an inversion between conditions and effects of the felicity of pornographic speech. Of course, in presenting women's silencing as following rather than preceding pornographic speech, it obscures alternative explanations for sexual subordination. This ambiguity allows Langton to present examples of sexual violence as evidence of pornographic silencing. Consider the following example, in which a woman attempts to refuse sexual advances: Sometimes "no," when spoken by a woman, does not count as the act of refusal. The hearer fails to recognize the utterance as a refusal; uptake is not secured. In saying "no" she may well intend to refuse. By saying "no" she intends to prevent sex, but she is far from doing as she intends. Since illocutionary force depends, in part, on uptake being secured, the woman fails to refuse. (Langton, 2009b, p. 54).
In this passage, just like in Marchiano's case, no actual evidence is provided to demonstrate that the uptake is not secured. On the contrary, this is simply stated as a matter of fact. Nevertheless, it is insufficient to exemplify a case of illocutionary disablement by exposing the speaker's interpretation of its own utterance, rather than the one of the hearer, who is supposed to be responsible for the silencing act. To this purpose, it would be necessary to explain how men misunderstood women's "No." By avoiding to explain this point, Langton describes men's role in the pornographic conversation by means of the very objectifying speech she formerly ascribed to pornographers: as a matter of fact, all the aforementioned examples involve a speaker and a character who is held unable to speak. But the latter one is not the woman, as she is portrayed by pornographic speech, nor is it the pornographer, even though he is not directly involved as a speaker. It is the man, whose interpretation of both women and pornography's speech acts is not deemed worth of being examined. Hence, just like pornography ranked women as objects, Langton ranks men as unable to recognise sexual violence.
By contrast, much information is provided about how the woman intended to act by enunciating her "No." Langton thus presents her description of silencing as objective, even though she actually privileges women's perspective on such a phenomenon. So, she ranks women as credible while testifying sexual violence. Eventually, pornographic speech is arguably denounced for its trumping women's freedom to express a reliable opinion about its effects. Therefore, Langton renegotiates the common conception of women's subordination by presenting it not as a mere disempowering circumstance, but rather as the very reason why the reader should adjust his epistemic stance in order to accept women's testimony about sexual violence. Indeed, she highlights the vantage point of silenced subjects on such a phenomenon.
After the publication of "Speech Acts and Unspeakable Acts," many authors benefited from the shift of perspective introduced by Langton, a turning point whose importance can be deduced by considering its relationship with alternative conceptions of pornography. Langton herself presents her position about pornography's authority as the core of an intense controversy against two other streams of thought: the former is represented by liberals, who recog-nise, like Langton, that pornography has authority, but define this authority as the one that all citizens of a democratic state have; the latter is attributed to postmodern philosophers, who, according to Langton, think that neither pornography nor women have any authority, because everyone's freedom of expression is limited by culture and language. These two perspectives are presented as a rigid binary, since accepting one clause implies refuting the other, even though they both reach the same conclusion: "If censorship is nowhere, women are not silenced at all…. If censorship is everywhere, it might as well be nowhere" (Langton, 2009a, p. 6).
Langton's method thus introduces a third alternative, the one according to which it is from the standpoint of women (and not of the state) as a subordinated group (and not as victims of an all-involving set of power relationships) that pornographic silencing can be identified in its specificity. In this sense, pornography does not only have authority, but has too much authority, because it deprives other social groups of their power and of their status as free and equal citizens. This is the actual reason why pornographic speech should be censored.
From this perspective, Langton's method arguably presents an innovative political stance which has stimulated many philosophical elaborations, but it may also have produced unintended and detrimental effects. In particular, her conception of the power relationship between pornography and women, as well as the substantive notion of authority that underlies it, has often led her to actively discredit direct political intervention by subordinated groups. Already in "Speech Acts and Unspeakable Acts," she questions the plausibility of using "the speech of competition to counter pornography's monopoly" by saying that "[a]ll this may be possible if women can indeed fight speech with more speech. But if pornography not only subordinates but silences women, it is not easy to see how there can be any such fight" (Langton, 2009b, p. 47). It is worthwhile to recall that arguments about plausibility and realism are generally based on contestable and unjustified presumptions about social reality: in this case, like in many others, Langton presumes that women's position of power is determined by pornography. She thus reifies the relationship between the two speakers and presents state intervention as the sole possible remedy to this structural oppression. Hence, it is now reasonable to ask: does Langton's argumentation aprioristically rule out and discourage the possibility of active resistance against pornography? If so, this consequence would have to be considered as a (perhaps unintended) side-effect of her performative reappropriation of epistemic authority. As a matter of fact, it is because she naturalises the idea that women's credibility depends on their being subordinated that the possibility of active contestation is presented as implausible.

Silencing the Allies: The Risks of Langton's Politics
In order to verify the aforementioned hypothesis, I will analyse the controversy between Langton and Butler. The relevance of this theoretical exchange for the present purposes lies in the fact that Butler is the sole post-structuralist philosopher with whom Langton entertained a durable discussion about the concept of power -traces of such a dialogue can be found also in her more recent works (2017; 2018). By analysing Langton's response to Butler's Excitable Speech, I thus aim to demonstrate that she superimposes her presumptions about social reality (in particular, her substantive conception of power) on her interlocutor's argumentation. Eventually, I will conclude that this partial interpretation entails presenting Butler's reasoning as implausible without providing any actual argument to support such a thesis.
Of course, this "credibility deficit" -to borrow a term from Fricker (2007, p. 17)-depends on the fact that Langton's substantive conception of authority impedes her to correctly interpret the notion of power that she labels as postmodern. As a matter of fact, she presents it by saying that "[c]ulture, discourse, perhaps language itself, all breath down your neck with their threats and orders… You are a prisoner in an invisible cage, invisibly gagged, allowed to say one thing, prevented from saying others" (Langton, 2009a, pp. 5-6). Considering that this formulation will later be used to criticise Butler's book, and that this book explicitly adopts a Foucauldian notion of power, it is convenient to notice that Foucault himself asserted exactly the opposite of what Langton contends. This is evident, for example, in the History of Sexuality, where he writes: "Power is everywhere; not because it embraces everything, but because it comes from everywhere" (1978, p. 93). Even more importantly, this thesis is the core of Butler's description of power, insofar as she presents it as a "shift from the subject of power to a set of practices in which power is actualized in its effects" (1997, p. 79). This of course means that culture and language shape the subjectivity of individuals, but it also implies that they are sensitive to all the variations of the practices in which power is embedded, so that their organisation can be modified by means of political activism. For this reason, the all-involving character of power cannot be understood as a constant limitation of individual will -it cannot be described as allowing or preventing individuals to say something, as if it could be experienced as a constriction. Rather, those philosophers inspired by Foucault contend that power conditions the very formation of individual self-conscience and desires, thereby shaping the social construction of reality which is prior to the possibility to choose what to say. Langton's misinterpretation is evidently based on a substantive conception of power, one that reifies "[c]ulture, discourse, perhaps language itself" as a hyper-subject who issues "threats and orders" (Langton, 2009a, p. 6), instead of describing them as contingent and mutable social practices.
Langton's reluctance to think power without subjectivity in turn explains her oversimplification of Butler's account of pornography. More specifically, since she presents power as something that can only be conceived as present or absent in preexisting subjects, she synthetises the content of Excitable Speech by saying that "Butler denies the authority [of pornography], and concludes that pornography does not subordinate women" (Langton, 2009c, p. 103). However, according to Langton's argument, this thesis is simply false, otherwise one should suppose "the normative beliefs of those exposed to pornography to be unchanged by it" (2009c, p. 110) But "there is evidence that consumers' normative beliefs are measurably altered following exposure to pornography" (2009c, p. 110). Hence, "[t]hese perlocutionary effects on normative beliefs are poorly explained by an assumption that pornography has no authority" (2009c, pp. 110-111).
This interpretation of Butler's thesis is questionable, because the argumentation of Excitable Speech does not straightforwardly deny that one's normative beliefs are influenced by pornography, nor does it assert that pornography lacks authority. If anything, it criticises the idea that such an authority can be deduced from what sexual practices pornography's consumers hold to be legitimate. Sure enough, Butler utterly criticises the argument through which [t]he elaborate institutional structures of… sexism are suddenly reduced to the scene of utterance [of pornographic speech], and utterance, no longer the sedimentation of prior institution and use, is invested with the power to establish and maintain the subordination of the group addressed. (1997, p. 80).
In order to criticise this assertion, Langton would thus have to interpret it as problematising the idea that an alteration of one's normative beliefs must depend on a previous illocutionary act of legitimation: indeed, in a society which is penetrated by institutionalised practices of subordination, one's convictions about what is legitimate can reasonably stem from various sources, not just from explicit acts of legitimation, and consequently not just -not even in the first place-from pornography.
The partiality of Langton's reading may depend on the fact that she understands Butler as following the same method she employed in "Speech Acts and Unspeakable Acts." Her observations clearly recall the thesis that "[p] ornography's effects may be best explained by supposing that it has the illocutionary force of subordination" (2009b, p. 48). By presenting normative beliefs as "perlocutionary effects" (2009c, p, 110), Langton thus obscures Butler's original thesis -i.e. that one's normative presumptions can be affected by something different than an illocutionary act of legitimation-and binds her to demonstrate that one's convictions about the legitimacy of sexual practices depend on a different illocution than pornographic speech. This argument could even be valid, if Langton had proven the illocutionary character of pornographic speech, but, as I pointed out in the first section, her argumentation falls into a circularity.
Far from highlighting a mere shift in the burden of proof, the fact that Langton displaces Butler's argumentation from its original context and judges it by reference to an evidence she did not herself provide should stress the political function of her discourse. This argumentative strategy proves to be even more efficacious, if one considers that it offers a merely evidential reading of Butler's book, one that radically excludes its interpretation as a performative. Hence, Langton's main political move consists in depoliticising Butler's discourse. As a matter of fact, as I mentioned above, every philosophical argumentation can be interpreted not just as a mere logical deduction, but also as a performative attempt to renegotiate common presumptions about social reality. This assumption can now be specified by means of a thesis argued by Bonnie Honig: insofar as every interpreter has to decide whether to value the performative or the descriptive dimension of the discourse she reads, by choosing the latter option, she necessarily obscures its political function (Honig, 2009).
As an example of this depoliticising effect of Langton's interpretation, consider how she understands the following passage: Through what means does the "as" turn into a "is," and is this the doing of pornography, or is it the doing of the very depiction of pornography that Mackinnon provides? For the "as" could also be read as "as if," "as if one were," which suggests that pornography neither represents nor constitutes what women are, but offers an allegory of masculine willfulness and feminine submission… one which repeatedly and anxiously rehearses its own unrealizability. (Butler, 1997, p. 68).
In Langton's parsing, this passage highlights "Mackinnon's grammatical mistake," that is, her confusion between the descriptive value of the verb "is" and the hypothetical meaning of the conjunction "as" (Langton, 2009c, p. 108). Thus, the proposition according to which as could also be read as as if is understood as highlighting the correct interpretation of pornography, rather than an alternative reading thereof. Langton's explanation of the passage as a constative is legitimate, but it is hardly compatible with other Butler's arguments and works.
As an example of this incompatibility, it is convenient to observe the passages of Excitable Speech in which the author criticises the following statement by MacKinnon: "Pornography makes the world a pornographic place through its… constructing the social reality of what a woman is" (MacKinnon, 1993, p. 25, in Butler, 1997. From Langton's perspective, Butler's doubts about the idea that pornography substitutes for a reality which "is figured as more original" (Butler, 1997, p. 66) should have a constative meaning. The question about what a woman is should therefore be answered by referring to another reality, which is even more original than MacKinnon's one.
But this presumption contradicts a thesis that Butler argues in other works, according to which women's identity is culturally constructed and cannot be objectively determined. Most importantly, consider Gender Trouble -a book which Langton quotes in her response to Excitable Speech-where she asserts that "the presumed universality and unity of the subject of feminism [i.e. women] is effectively undermined by the constraints of the representational discourse in which it functions" (Butler, 1990, p. 6). Similarly, by following Langton's interpretation, it would be difficult to explain Butler's proposal to "contest the very reifications of gender and identity" by assuming the "variable construction of identity as both a methodological and normative prerequisite" (Butler, 1990, pp. 7-8).
Hence, how else could Butler's text be interpreted? What potentialities does Langton obscure by reading it as a constative? Most plausibly, her interpretation conceals its political meaning, the idea that the allegorical reading of pornography is a legitimate alternative, but not a more faithful interpretation of the phenomenon in question. In this sense, Butler's argumentation may be read as a performative, one that suggests that nothing compels us to consider pornography as determining women's social identity -nothing but "the very depiction of pornography that MacKinnon provides" (Butler, 1997, p. 68). It may thus present the question whether pornography constitutes or causes subordination as undecidable, i.e. as a matter which cannot be objectively or impartially described. From this perspective, alternative interpretations of the phenomenon in question cannot be evaluated through standards like truth and coherence, but rather by reference to their political consequences.
Butler actually refers to such consequences while criticising the depiction of pornography and hate speech as illocutions. This is the core of her analysis of Langton and MacKinnon's legal proposals, according to which, since discriminatory language constitutes subordination, it has to be contrasted through the state's intervention -in particular, through censorship. In Butler's parsing, "writings that favour of hate speech legislation… minimize the possibility of a misappropriation by the law in favour of a view of the law as politically neutral" (1997, p. 98). More specifically, they obscure the possibility that even state-sanctioned speech can produce effects of subordination. 15 These passages have been overlooked by Langton, most probably because they cannot be evaluated through the evidential standards of a constative interpretation of Excitable Speech. Nevertheless, they are crucial for a performative interpretation of Butler's text, since they ground the conclusion that her criticism against the illocutionary conception of hate speech, as well as her advocacy for a perlocutionary explanation thereof, are motivated by a specific political end. According to my argumentation, this end consists in encouraging the spontaneous political initiative of subordinated groups to contrast oppression, namely by "decontextualizing and recontextualizing such [discriminatory] terms through radical acts of public misappropriation" (Butler, 1997, p. 100).
Such a political proposal is not thoroughly discussed by Langton, who by contrast chooses to reject it as implausible: in her parsing, "appropriation or deconstruction of their [of hate speakers] speech by parties who would otherwise be injured by it" simply is an "optimistic hope" (2009c, p. 114). As for the rest, no actual argument is provided to ground this accuse. Nevertheless, Langton could not formulate such an unjustified rejection, if she were not confident that her assertion would have been unproblematically approved by her readers. What is, then, the presupposition that makes her so confident? Most probably, it is the presumption that, if women are injured by pornographic speech, they are unlikely to contrast it on their own. If they lack the authority to engage in such a struggle, one can reasonably doubt that they will succeed in altering the current distribution of social power. It is therefore Langton's substantive conception of authority, i.e. the presumption that who lacks social power is simultaneously unable to act politically, that motivates her aprioristic rejection of appropriative politics.
To be sure, the argumentation I exposed in the present section is not a refutation of Langton's response to Butler. By considering it as a performative, I rather exclude the possibility that it is erroneous -in fact, it displays no patent inconsequence-and make room for the possibility that it is just partial. In this sense, Langton's unjustified refutation of appropriative politics can be explained by assuming that she expects her substantive notion of power to be shared by the majority of her readers. This in turn means contributing to naturalise a contingent (because actually unjustifiable) presumption about social reality and depriving all the alternative conceptions of power of their credibility. One may name plenty of texts which evidence this tendency: just to limit the examples to Butler's critics, consider Martha Nussbaum's accusation against "the fatalistic idea that we are prisoners of an all-enveloping structure of power" (1999), or Lisa Schwartzman's thesis according to which Butler neglects the importance of existing power relationships for her very political proposals (2002). By joining such a trend, Langton reproduces the conditions to silence advocates of alternative emancipatory strategies. And of course, this political move cannot be deemed deplorable by reference to universal normative standards, but has the cost to legitimate unjustified scepticism toward contestatory politics.
This legitimation can be considered as another illocution performed by Langton's text, beside the performative reappropriation of epistemic authority highlighted in the previous section. However, one may ask, does this reappropriation necessarily occur at the expense of the credibility of other theoretical and political movements? Is it possible to conceive subordination without implying inability to act politically? 15 Butler provides various instances of such a "misappropriation by the law" (1997, p. 98). See Butler, (1997, pp. 97-8): "[T]he state not only constrains speech, but in the very act of constraining, produces legally consequential speech: not only does the state curb homosexual speech, but produces as well--through its decisions--a public notion of the self-censoring homosexual; similarly, it produces a public picture of obscene black sexuality, even as it claims to be curbing obscenity; and it produces the burning cross as an emblem of intelligible and protected speech."

Conclusive Remarks
Questions like those I have raised in the end of the previous section deserve a more detailed account than the one I can provide in the present article. Nevertheless, it may be useful to sketch a methodological proposal which is directly related with the themes I treated so far, namely the influence of current conceptions of power on the distribution of credibility among alternative accounts of political agency. In particular, since I assumed a Foucauldian conception of power in order to analyse Langton's texts, it is now convenient to explain how such a presumption, if employed in the interpretation of political argumentations, can avoid the problems that emerge from a substantive conception of power.
Most evidently, the Foucauldian perspective does not deny the existence of current power relationships, as some authors have contended. Locating power in practices does not mean that social hierarchies randomly change, nor does it entail that this change is a reason to abstract from the perceived effects of domination. It rather implies that the linguistic categories and the normative concepts that qualify some social groups as subordinated exercise a conservative effect on emancipatory politics, just like I tried to demonstrate in Langton's case. Hence, social hierarchies exist, but they are not a sufficient theoretical paradigm to interpret the way domination works. Often power operates through conventional practices which reenforce the subordinated status of some groups by obscuring their political potentialities, and this process involves subjects who are not qualifiable as powerful or powerless (Young, 2011).
Tracing such a linguistic and cultural conditioning can thus provide some useful inputs to problematise the concept of epistemic injustice. So, consider the question: does the controversy between Langton and Butler display a form of epistemic injustice? One may suppose that some parameters are satisfied in this case: for instance, Butler could be said not to be afforded sufficient credibility in comparison with the evidence and arguments she provides. 16 As a matter of fact, Langton does not even mention Butler's analysis of Anita Hill's testimony, which is the main evidence for her thesis that sexism and racism cannot be reduced to the utterance of hate speech (see Butler, 1997, pp. 82-86). Furthermore, Butler's credibility is arguably undermined because of her being a "postmodern" philosopher, who endorses a defeatist conception of power (Langton, 2009a, pp. 5-6;2009c, p. 103), so that she can be said to suffer some sort of identity prejudice. 17 Finally, Langton has an obligation to match her credibility judgement to the evidence and arguments provided by Butler, since the conversation they entertain occurs in a scientific context. 18 Is this sufficient to determine this case as testimonial injustice? The answer is arguably uncertain. Indeed, even if Butler's credibility deficit depends on a partial interpretation of the evidence she provides, it is not (just) her statements about a matter of fact that are silenced. If anything, one may say that a merely evidential reading of her text obscures the political proposals she advances to contrast hate speech. As a result, even if all the evidence she presented in Excitable Speech were accurately analysed by her audience, some sort of misunderstanding would have anyway occurred.
One may thus conclude that epistemic injustice is not involved in this process: perhaps, the credibility deficit in question depends on Langton's attributing a different performative meaning to Butler's utterances than the one she intended. However, I would not resort to the category of intentionality in my argumentation, because, as I have already mentioned, it seems difficult to empirically individuate it. Furthermore, since I assume that a philosophical text can be understood in various ways (at the very least, as a constative and as a performative), I also make room for the possibility that alternative interpretations can be likewise legitimate, no matter how incompatible they are. More specifically, I presume that no interpreter can totalise all the possible implications of a text, and that it is impossible to unambiguously reconstruct the situation that constitutes its context. In adopting this methodological perspective, I do not simply recognise that interpretation is necessarily partial and selective, but attribute a constructive function to this activity. The reading practice does not merely entail passive reception of a preexisting meaning, but also requires a projection of contextual presumptions on the text (Butler, 1997;Derrida, 1998). In the case of my analysis of Langton's response to Butler, such a projection involves presumptions about the nature of social reality. But, even for this reason, I cannot rank Langton's reading as illegitimate or incorrect without superimposing my own social-ontological presumptions on it -except in case of patent logical errors. Hence, I cannot admit the possibility that Langton actually silenced Butler, since I cannot resort to universal standards to evaluate her interpretation. If anything, lack of credibility on the part of Butler is a more appropriate notion to describe the case at hand.
Anyway, this form of credibility is not directly related to testimony, since it does not concern the reliability of Butler's statements about a matter of fact, such as past events or general truths. In other words, it is not to be intended just as epistemic authority, but also as practical authority: it is not the authority of a speaker qua knower, but rather the authority of a speaker as a political subject, one who is held reliable while elaborating programs for collective action. In this case, evidence required to evaluate this utterance are (at least partially) impossible to collect, since they do not refer to something which is already happened, but to an expectation about future outcomes of somebody's actions (Derrida, 1992). 16 On the importance of this criterium, see e.g. Fricker, 2007, p. 19: "Epistemological nuance aside, the hearer's obligation is obvious: she must match the level of credibility she attributes to her interlocutor to the evidence that he is offering the truth." How is it possible to trace back this alternative form of epistemic injustice? As I have suggested, drawing on Bonnie Honig's works, one way consists in individuating constative interpretations of political texts. Indeed, these obscure and neutralise their performative function, i.e. proposing political strategies or renegotiating current assumptions about social reality. Of course, this implies reconstructing the context of the interpretation in question by individuating the social ontology which it tends to reproduce along with the problematic political consequences thereof. In this way, it is possible to avoid the risk to reproduce the phenomenon of epistemic injustice one tries to prevent (that is, the risk to reduce a text's constative interpretation to another constative utterance), because assuming that a text naturalises a contingent social ontology means recognising its political meaning. As a matter of fact, such a naturalisation is not conceivable unless it is interpreted as a performative, and, more specifically, as an exercitive or a verdictive illocution.
By leaving room for this possibility, I also reject Butler's idea that hate speech has to be conceived as an instance of perlocution. According to my arguments, interpreting phenomena such as epistemic injustice or hate speech as illocutionary acts does not entail reproducing the subordinated status of some social groups. On the contrary, the very hypothesis that a discourse is depoliticised by a constative interpretation implies the obligation to provide an alternative reading of the depoliticised text, one that highlights its political potentialities. Accepting this burden of proof is ultimately what qualifies my account of epistemic injustice as agonistic, since it refers to a credibility deficit which cannot be contrasted by referring to empirical evidence, but rather through the performative practice of renegotiating discriminatory social-ontological presumptions.