Pornography and speech acts

Critics typically respond to anti-porn feminist arguments (according to which porn eroticizes the degradation of women; celebrates violence against them; portrays a false depiction of female pleasure and sees it as ancillary to male pleasure; etc.)Footnote ⁴ by arguing that pornography is a form of speech or expression, and should thus only be restricted when it directly and demonstrably causes harm (Mill Reference Mill1978).Footnote ⁵ Now, whilst some have sought to establish causal links of this kind (for instance, between porn and rape), the evidence on this matter is shaky at best.Footnote ⁶ Against this background, MacKinnon has pioneered an altogether different brand of anti-pornography argument, claiming that porn directly constitutes the wrongful subordination and silencing of women (1987, 1993).Footnote ⁷ In particular, if pornography silences women, letting pornographers speak might curtail the free speech of women—thus turning the freedom of speech argument on its head. Whilst this claim has been dismissed as a “sleight of hand” in court (American Booksellers Inc. v Hudnut 1984, 598 F.Supp. 1316), Langton has insightfully used speech-act theory to defend its coherence (1993) and relied on some empirical evidence to support its plausibility (2017).

According to the theory of speech acts (Austin Reference Austin1962), speaking always involves three dimensions. When I tell you “You are invited to my wedding!” I use the locution constituted by those six words to attempt to perform a specific speech act or illocution—in this case, an invitation; the perlocutionary dimension, in turn, is the effect which the performance of an illocution produces—for instance, you coming to my wedding, or becoming jealous. For illocutions to succeed, certain “felicity conditions” must be in place—some of which pertain to the speaker, others to the hearer, others still to the context and background norms. For instance, you might not understand that I am genuinely inviting you to my wedding because I constantly make jokes on the matter; or my speech may not count as an invitation if relevant norms dictate that invitations may only be issued by the parents of the betrothed. Finally, a speech act might succeed as such, yet fail in its intended perlocutionary aim—you might understand yet decline my invitation.

Within this paradigm, it becomes obviously inaccurate to claim that speech—unlike conduct—is something which one, bar exceptional circumstances, has always a claim to engage in (and whose performance should therefore be legally protected). Clearly, free speech defenders do not contend that all speakers should nearly always be able to, say, marry people or declare them guilty of murder. What they must mean is, instead, that the performance of certain, specific categories of speech acts—e.g., expressions of opinions, ideas, artistic concepts, etc.—should enjoy said special protection.Footnote ⁸ Thus, MacKinnon's startling claim could be reformulated as such: when certain conditions are in place, the speech of pornographers constitutes a different kind of speech act, not covered by those categories. Some things we can do with our words are particularly striking. If I occupy a certain role, if the two people in front of me meet certain criteria, and if certain other background conditions hold, my saying “I hereby declare you wife and wife” actually marries them. Speech acts of this kind are often called performatives, and interrogating their correct moral and legal status is clearly legitimate.

In Langton's view, the best interpretation of MacKinnon's apparently mysterious claim is that she had something like performatives in mind. Two categories of performatives are particularly relevant here: (1) verdictives, which constitute “authoritative deliver[ies] of a finding about some matters of fact or value” (Langton Reference Langton1993, 305); and (2) exercitives, which confer powers and rights on people, or deprive them thereof (Langton Reference Langton1993, 304). Both can, in Langton's view, subordinate as well as silence. Verdictives can subordinate people, by authoritatively ranking them as inferior; and silence them, by establishing that their speech acts ought to be ignored, or interpreted in ways that are systematically different from what the speaker actually means. For instance, a Soviet doctor diagnosing a political dissenter with “sluggish schizophrenia” under Stalin's regime performs a verdictive that arguably silences the dissenter's speech—which will now be, by default, interpreted as the delirium of a mentally ill individual rather than a genuine expression of dissent. Similarly, exercitives can deprive people of certain rights, and declare them incapable of performing certain acts with their speech. Langton argues, for instance, that a legislator in Apartheid South Africa declaring “Blacks are not permitted to vote” subordinates black citizens, by ranking them as inferiors; and silences them, by depriving them of the power to perform the speech act of voting.

Whilst the sense in which performatives can subordinate is probably fairly clear, it may be worth clarifying in which way they can silence. Ordinary language most often understands silencing in a locutionary sense, that is, as actually preventing people from speaking, period. Langton argues, however, that the key to understand MacKinnon's claims is illocutionary silencing. Her initial definition of the latter is that it consists in depriving speakers of the capacity, not to speak, but to perform certain speech acts: “Let them speak. Let them say whatever they like […], but stop that speech from counting as an action. More precisely, stop it from counting as the action it was intended to be” (Langton Reference Langton1993, 299; emphasis in the original). Unlike this general characterization, however, virtually all of Langton's concrete examples are cases where the attempted speech of speakers suffers lack of “uptake” (Austin Reference Austin1962; Langton Reference Langton1993) in particular, that is, where hearers problematically (for reasons other than a random misunderstanding) do not understand what speech act the speaker is trying to perform. This is somewhat unfortunate, and scholars have rightly pointed out that plausible candidates of illocutionary silencing can occur in ways other than via failed uptake. Hearers can, for instance, understand which speech act the speaker is trying to perform, yet unfairly doubt their sincerity, seriousness, or reliability (McGowan Reference McGowan2003, Reference McGowan2014; Caponetto Reference Caponetto2017, Reference Caponetto2021). When this occurs because of porn, we might say that women suffer “testimonial injustice” (Fricker Reference Fricker2007, 17–18)Footnote ⁹—a kind of epistemic injustice which occurs when a speaker receives an unfair deficit of credibility from a hearer. This can happen, say, when women are taken to be either positively lying when they say “no,” or “confused” about what they actually mean and want. Alternatively, silencing can happen when certain speakers are taken not to be entitled to perform certain speech acts to begin with—for instance, to have no standing to refuse sex (Caponetto Reference Caponetto2017; Hesni Reference Hesni2018). These insights have led Caponetto to offer a novel definition of illocutionary silencing, whereby the latter occurs when a hearer systematically fails to recognize the fact that a given speaker meets certain relevant criteria to perform certain speech acts (this covers uptake but also sincerity, seriousness, entitlement, etc.). Whilst I agree that we should go beyond uptake and that Caponetto's definition is an insightful step in the right direction, I think we are better off staying somewhat closer to Langton's original characterization. Caponetto's definition, whilst enabling us to go beyond failed uptake, appears under-inclusive in other respects. Take, for instance, a legal system where women are not deemed competent to give evidence in court. Judges following this law are not failing to recognize anything, in their capacity qua judges; yet, having a definition of illocutionary silencing which does not enable to call this a case of silencing appears a great price to pay indeed. Of course, one might say that the law itself is committing a recognitional mistake here; but then it is not quite the hearer, as in Caponetto's definition, who is committing a recognitional failure. I submit that a more straightforward way of capturing the same idea is to say that one is silenced when one is systematically deprived of the capacity to perform certain speech acts that one has a rightful claim to perform.Footnote ¹⁰ This has the same virtues as Caponetto's definition, in that (1) it goes beyond uptake and (2) allows for the possibility that one may justly suffer illocutionary disablement (and thus not be silenced).Footnote ¹¹ At the same time, it avoids the implausible implications which the criterion of recognitional failure leads to. I do endorse, however, Caponetto's emphasis on systematicity, although I shall understand this notion in more common-sense terms:Footnote ¹² speakers are silenced when they are incapable of performing certain speech acts which they ought to be able to perform with a certain level of consistency and predictability—at least vis-à-vis some salient audiences—and not just in the odd random case.Footnote ¹³

A crucial aspect of illocutionary silencing is that one can be silenced with regard to speech acts other than those which traditionally pertain to freedom of speech. Hence, illocutionary silencing comprises both, say, being systematically misunderstood in one's attempts to express a view, but also being deemed unfit to make certain exercitives (e.g., “I do not consent to this”). I believe this to be the correct outcome, if perhaps not perfectly in line with ordinary language, because it allows us to include all cases where one is harmed, through speech acts, in one's capacity to perform certain speech acts.Footnote ¹⁴ Finally, it is worth noting that silencing can occur as a result of subordination: if women are established as mere playthings during sex, then their withdrawal of consent does not matter, even when expressed (and understood).

We can now appreciate Langton's point: what if MacKinnon had illocutionary silencing and subordination in mind? For that argument to go through, pornographers must be able to perform subordinating and silencing verdictives and/or exercitives. One crucial felicity condition is necessary for both: speakers must have authority within the relevant domain. If pornographers are a shunned, powerless fringe, then their speech cannot subordinate or silence; if, however, they are recognized authorities in the domain of sex for certain audiences, then their speech positively sets standards. If this is the case, pornographers are able to perform both verdictives (authoritatively establish truths about sex) and exercitives (establish “which moves in the sexual game are legitimate,” Langton Reference Langton1993, 312). If the verdictives convey, say, that women's resistance to sexual advances is just another form of foreplay; and if the exercitives establish, say, that certain fantasies are not just “hot,” but also constitute an appropriate way to treat women; then porn can indeed both silence and subordinate women. Langton clarifies how this sheds light on MacKinnon's use of Linda Marchiano's case (1993, 321). In publishing her memoir Ordeal, ex-porn worker Marchiano intended to denounce her fate as a slave in the porn industry; the book, however, was advertised as an “erotic autobiography” when it first appeared; her story, coming from her, could not but be seen as a titillating piece of erotic literature. If an episode like this is not just a freak, extremely unfortunate event, but occurs because porn is systematically perceived as setting standards on what women desire (at least for certain salient audiences), then the claim that pornography subordinates and silences women is perfectly plausible—although whether this is actually the case depends, of course, on empirical factors.Footnote ¹⁵

Publishing porn and the epistemic, informal authority of pornographers

Langton's argument has been criticized for being itself a “sleight of hand”: several critics, in particular, have argued that she has failed to show how porn can constitute illocutionary subordination and silencing (Jacobson Reference Jacobson1995; Butler Reference Butler1997; Bird Reference Bird2002). Many compelling responses already exist (Hornsby and Langton Reference Hornsby and Langton1998; McGowan Reference McGowan2003; Mikkola Reference Mikkola2011; McGowan et al. Reference McGowan2011). One further challenge to that effect, however, remains largely unmet (Saul Reference Saul2006);Footnote ¹⁶ addressing it, I believe, is not only possible, but also indirectly brings to the fore some specific features of the kind of authority possibly enjoyed by pornographers, which shall become relevant later.

Saul argues that Langton makes a category mistake: pornographic material, as such, is merely a locutionary medium. To illustrate this, she introduces the example of Ethel, who lives in an environment where people communicate to each other only by exchanging notes. Ethel writes “I do” on a piece of paper and “has a very eventful week, during which she uses the sign to get married, to agree to return her books on time, and to confess to murder” (Saul Reference Saul2006, 235). Now, surely the sign as such does not constitute any of these illocutionary acts—it is merely the locutionary medium used to perform all three. Illocutions are utterances in context; this is the whole point of distinguishing them from locutions, after all. Similarly, a piece of pornographic material as such is no illocution whatsoever. We therefore must ask what the relevant “utterance in context” is in this case: is it the production of the pornographic material or its consumption (say, a viewing)? Surely the latter, Saul argues. The actual production of pornographic material is like Ethel's writing of her sign: it is nothing but the creation of a locutionary tool, which can then be used to perform different illocutions. However, depending on context, viewings may or may not be subordinating and silencing; a radical feminist, after all, can successfully show pornographic material in a lecture in order to vindicate MacKinnon's argument. Therefore, claiming that pornography as such subordinates and silences women is, indeed, a sleight of hand after all.

Saul is right in drawing our attention to utterances in contexts. Viewings, however, cannot be the right category of utterances in context, in that MacKinnon and Langton clearly have in mind speech acts performed by pornographers (whereby the latter term is understood widely, as including writers, directors, producers). A viewing of, say, a porn video is, instead, performed by whomever organizes the viewing, rather than necessarily by the pornographer who wrote/directed/produced it. However, Saul has neglected a third possibility: the publication of pornographic material.Footnote ¹⁷ Publishing pornography (in whatever form) is indeed an utterance in context, just like publishing an op-ed in a newspaper. Unlike Ethel's neutral, multi-purpose note, it means making a meaningful statement. When we publish something, we publicly put content “out there” via specific channels and within a given context—which jointly co-determine (together with the actual content of what we publish, certain features about ourselves, and certain features and predispositions of hearers and viewers—in sum, jointly with the broad set of felicity conditions) the nature of the specific illocution we thereby perform. This is equally true of a journalist, a vlogger, the author of a book, and the producer of a porn movie: it is one thing to conceive and produce the material, another is to put it out there.

Now, saying that we should compare the publishing of pornography to speech acts such as the publication of op-eds seems to put pornography back into the protected category of expressions of opinions and ideas. Publishing something does not, however, always and necessarily constitute the speech act of expressing a view in the free market of ideas. When the American Psychiatric Association (APA) publishes the latest edition of the Diagnostic and Statistical Manual of Mental Disorders (one of the standard references on mental conditions for the discipline) it does not only express a view—it sets a standard. Publishing an edition where homosexuality no longer features does not mean, for instance, expressing the “opinion” that gay people are not mentally ill; it means that, from now on, not considering homosexuality an illness becomes the new, authoritative default—in a way that directly informs clinical practices and diagnostic criteria. Publishing porn may do something similar in the domain of sex—at least for some relevant and salient audiences. Publishing mainstream porn content can count, for certain salient audiences and given certain pre-conditions, as the “putting out there” of authoritative content—which, say, instructs said audiences about how women are to be treated and understood during sex.Footnote ¹⁸ This can occur, for instance, if it is the case that vast amounts or salient groups of teenagers see porn not only as a tool for sexual arousal, but also as the primary “go to” place to find out what sex is like.

An act of publishing can fail or succeed as an illocution and/or in its intended perlocutionary aims, just like any other speech act. I might want to persuade many people by publishing a pamphlet, and succeed in the illocution of putting out there a forceful opinion, yet suffer perlocutionary frustration in that most people are not persuaded. Or I might intend to exercise authority (rather than just express a view) by publishing a manual, yet audiences may not take it as a canon—thus causing my intended illocution to fail. For some speech-act theorists, the opposite can also occur: one can sometimes perform a speech act that one does not intend to perform. In the 1979 film Monty Python's Life of Brian, crowds elect Brian of Nazareth as their chosen religious guru—as a result, everything he says, despite his intentions, counts as authoritative religious teaching for them (including his desperate plea “You are all individuals!”, whereby he actually intends to reject the very role of spiritual leader). Similarly, when pornographers publish porn, this can arguably count as setting standards on sex for some audiences (if the relevant felicity conditions obtain) even if unintended. Pornographers might simply be exploiting the opportunities offered by the web, by circulating more and more material in order to make more and more money, following “trendy” tropes (e.g., rape fantasies). Indeed, with these perlocutionary aims in mind, they might not even give much thought to which illocutions they might be performing at all—they might not even intend to “express” certain views (i.e., to make truth-apt statements about sex), let alone to exercise authority. They might just be in the business of filming certain kinds of scenes again and again, because they sell well. However, the thereby arising omnipresence of pornographic material—together with, say, prevailing sexist normsFootnote ¹⁹ in society and the absence of alternative sources of sex education—can lead, say, the current generation of teenagers to see porn as a “go to” place to learn about sex, and thus to take certain tropes, chosen by pornographers because they sell well, as truths about women. On this reading, a speech act can be an exercise of authority even if its performer does not intend it to be; thus, hearers can not only ratify whether the intended speech act is successfully performed or not, but contribute to determining which speech act it is (see, e.g., Langton Reference Langton, Gardner, Green and Leiter2018a and MacDonald Reference MacDonald2021). This account is contested in the literature, and I do not have the space to adequately defend it here. The only remark I shall make here is that at least a weak version of this view, according to which hearers have the power to determine the nature of the speech act performed at least in some cases, is necessary to make sense of some important everyday concepts—especially the idea that those who have authority should watch their words, because, coming from them, they are not “just words.” The idea that someone with authority might exercise that very authority even when they do not intend to, just because of the privileged position they occupy, is a powerful intuition, which we should be able to account for (I owe this to Kukla Reference Kukla2014). This is the way in which, if Langton is right, pornographers might come to hold authority.Footnote ²⁰

Now, the publication of something can count as an authoritative act for some audiences and not for others; what is more, its authority can also be less than perfect—that is, it is possible to make a dent in it. These are two ways in which authorities can be partial. For instance, it could be that porn is only taken as authoritative by current teenagers—who are simultaneously overexposed to it, largely illiterate about sex, and not exposed to sufficient good-quality sex education. For this salient audience, porn might constitute not only a source of sexual arousal, but also a “go to” place to teach oneself about sex, in a way that teenage magazines might have done for previous generations—whilst for many other groups porn has no authority whatsoever. Of course, only empirical evidence can tell us whether porn really has authority of this kind, for whom, and how much of it (how robust the authority is). The kind of evidence we need is not, however, on whether “porn causes rape,” but rather on this very authoritative status. In the series Mrs Fletcher, for instance, the main character's 18-year-old son is genuinely taken aback when he realizes that certain “moves” he has learnt from porn—and “learnt” really is the right word here—are perceived as forms of assaults by one of his sexual partners. What needs to be settled empirically is whether this is a plausible scenario (for some evidence, see Maddy et al. Reference Maddy, Liz Kelly, Garner and Kanyeredzi2013; Langton Reference Langton and Mikkola2017, 28-35; Doornwaard et al. Reference Doornwaard, den Boer, Vanwesenbeeck, van Nijnatten, ter Bogt and van den Eijnden2017). In most cases of this kind, authority will also be partial in that it is possible to erode it: we can imagine Mrs Fletcher's son starting to question whether he really should have relied on the authority of porn so much; but we can also imagine cases where instilling doubt might be very difficult, even after experiences of this kind.

This response to Saul enables us to zoom in on two crucial features of the authority of pornography. First, like the APA and unlike a legislator, its authority is epistemic in kind.Footnote ²¹ Recognizing an entity as an epistemic authority means being committed to defer one's judgment to it on relevant matters of knowledge—what it says counts as standard-setting for us. This constitutes genuine harm if the epistemic authority is held without good reason, and especially so if it is wrongfully used to rank certain people as inferior, irrational, mentally ill, etc. Epistemic authorities can make a major practical difference. When an epistemic authority declares certain people mentally ill, for instance, it thereby has the power to sanction as appropriate certain behaviours that would otherwise count as disrespectful and indeed often impermissible (i.e., involuntary psychiatric treatment). If pornographers have authority, it is authority of this kind, mutatis mutandis: they can have, for instance, the authority to establish that “no's” uttered by women can be ignored, or that women may permissibly be used as playthings during sex.

This is compatible with epistemic authorities being challenged, and with their power being imperfect. Dissenting voices to mainstream psychiatry exist, and can at times get some traction—but this, up to a point, does not result in a full loss of authority for the medical establishment. Langton herself notes that the success of illocutions is not a sharp matter of yes or no, but a question of degree (Reference Langton1993, 301 and 308–09). What matters is for the relevant speaker to enjoy authority in a sufficiently robust and systematic way. In other words, there must be significant and salient audiences, and not just the odd case here and there, who believe certain things about women and sex because this is what porn shows. This is why a given content can be consumed in different ways by different audiences and still be meaningfully authoritative. If porn counts as authoritative for certain salient audiences, this will of course affect future acts of consumption, which will be disproportionately taken to be acts of learning. However, just like the APA's manual can be read with critical or even derisory aims in a conference of heterodox psychologists without thereby losing its authoritative status, similarly porn can sometimes be used for different purposes (including for feminist advocacy). As we shall see below, the fact that the authority of wrongfully authoritative speakers can be partial and imperfect is one of the reasons why there are more than one way to dismantle it.

A second, crucial feature of porn's authority is that, if it exists, it must be informal. No official legal construct confers authority upon pornography—feminists would say that what upholds it is the ideology of the patriarchy; others might say that pornography holds authority because it is, currently, the dominant or even only game in town for teenagers who want to find out how sex works. Now, can informal authorities (whether practical or epistemic) subordinate and silence? For much speech-act theory, the answer is clearly yes. The law is one social construct among others, if a very important one (see, e.g., Searle Reference Searle1995). Thus, agents can come to enjoy authority through de facto, informal means without having it formally conferred upon them.Footnote ²² This is true of both practical and epistemic authorities. A criminal or terrorist organization (say, the Camorra or Hezbollah) may function as an ersatz state in certain circumstances. Similarly, whereas some epistemic authorities are officially sanctioned, others function through informal mechanisms. A particularly congenial example is the case of Pickup Artists (PUAs), a community of self-proclaimed seduction experts who have gained a lot of traction in recent years. A common trope among PUAs is the idea of “last minute resistance,” according to which women often “resist” sexual intercourse just before it is about to happen (even if they are actually interested) because they are either biologically or socially “hardwired” to do it (Urban Dictionary n.d.). If salient groups of men come to see PUAs as genuine experts (they do, after all, publish manuals and teach workshops), their claims will become epistemically authoritative for them—for certain audiences, “last minute resistance” will be a serious idea, and women resisting sex after having initially showed interest will no longer count as having genuine second thoughts. Similarly, and in much larger numbers indeed, pornography may come to enjoy informal authority over teenagers. One may say that, by allowing porn to be omnipresent; by not providing alternative forms of sexual education with sufficient impetum and thoroughness; and by doing so in the knowledge that we live in societies that are already ridden with sexism, we enable pornography to come to hold informal epistemic authority. We send the message that “this is how men are taught, around here […] about how women are to be treated” (Waldron Reference Waldron2012, 91; emphasis in the original—also quoted in Langton Reference Langton and Mikkola2017, 34). Now, as Langton argues (Reference Langton, Fogal, Harris and Moss2018b), speakers can hold epistemic authority because they are experts, or because they enjoy credibility qua experts without being actual experts—when the latter is the case, the authority is held without good reason: the speaker counts as epistemically authoritative, but should not. Arguably, both PUAs and porn fall in this category.

Summing up, a successful response to Saul can be offered—and it allows us to draw attention to certain underappreciated features of the authority of porn: its being epistemic rather than practical; informal rather than formal; and partial/imperfect rather than absolute. These features will turn out to have important implications.

Felicity conditions, not locutions! On what exactly is inadmissible about inadmissible illocutions

The claim that (publishing) pornography subordinates and silences women is therefore perfectly coherent. If it is also empirically correct, Langton argues, this might cast doubt on whether pornography should enjoy the special protection granted to free speech: “Women wish to be able to speak some important action […]. 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” (Langton Reference Langton1993, 330, emphasis added). This section argues that, contrary to what this quotation might suggest, the speech-act argument against pornography is not a game-changer for the status of censorship, even if successful in and of itself.

One clarification is necessary before I proceed. Langton carefully shies away from actively vindicating censorship (Reference Langton1993, 299–300, 313–14, 328).Footnote ²³ Her argument, however, has been widely interpreted to provide some positive support for it (Jacobson Reference Jacobson1995, 65 calls it a “new argument for censorship”; but see also West Reference West and Zalta2004), or at least to lift the strong presumption against it (see, among others, Jacobson Reference Jacobson1995, Reference Jacobson2001; Bird Reference Bird2002). If pornographic speech constitutes harm in and of itself, so the story goes, censorship becomes at least prima facie an option, if not necessarily all things considered recommended—say, because measures short of censorship prove to be sufficient, or because censorship might dangerously drive the porn industry underground in a way that might harm many women in it. Thus, Langton makes a very careful claim indeed, but many commentators take such careful claim to imply at least that the speech-act argument against porn is a game-changer, in that it puts the issue of censorship on a different footing. We can at least, it is suggested, entertain the issue of whether to censor porn or not. My claim is that the speech-act argument is indeed ground-breaking in showing that porn can be constitutive of harm but involves no shift of perspective with respect to the status of censorship.

The argument is very straightforward but might appear counter-intuitive; I shall therefore first lay it out, and then offer some examples to support it. If an act is wrongfully harmful, it becomes justifiable to (try to) counter it—indeed, it might even be required to do so. And if we have established that a specific harm is committed with words, it is tempting to conclude that it becomes prima facie (if not necessarily all things considered) justifiable to stop those words from being uttered to begin with. Censorship, in other words, appears to be allowed back into the policy toolkit—although whether it is also, ultimately, the right thing to do remains to be seen. This is not, however, what the speech-act argument establishes. The problem it identifies starts and ends at the illocutionary level: it is wrong for the speech of pornographers to count as an authoritative, subordinating, and silencing act, not for such speech to occur at all. Thus, suggesting that the strong protection granted to free speech might not apply to porn rests on a double standard. When it comes to understanding what (publishing) pornography is, it is pointed out that it is a kind of illocution; when it comes to deciding what to do about it, one reverts back to putting the emphasis on its locutionary dimension.

Recall that MacKinnon's argument and its speech-act reconstruction claim to be fully compatible with, and indeed grounded in, the principle of freedom of speech: “It is the same social goal, just other people” (1987, 156). Banning locutions, however, means disallowing all illocutions that need those locutionary means to be performed (included expressions of opinions and ideas, however abhorrent these may be). Jacobson (Reference Jacobson1995) unpacks this thought by claiming that freedom of speech should be understood as freedom of locution. Is he right? Not quite, yet there is a kernel of truth in this claim. Freedom of speech, as we have seen, is the protected freedom to perform a certain category of illocutions; however, it is true that, for that to be possible, locutions to perform those speech acts must remain available. If, say, I wish to express a view, but this is taken as an authoritative verdict by certain salient audiences (and it should not be), it is the latter issue that constitutes a problem and should be fixed. This does not mean downplaying the speech-act argument in any way: establishing that (publishing) pornography may very well constitute harm is ground-breaking and its importance cannot be overstated. If empirical evidence lends support to the view that the problem exists and is extensive, this potentially grounds a case for trying to dismantle the misplaced authority of porn (and the felicity conditions that sustain it) very forcefully indeed. What pornographers must suffer, however, is not censorship, but rather (just) illocutionary disablement: they must be deprived of their power to silence and subordinate.

This claim might seem striking; yet, at closer look, it captures the standard way in which we already tackle other kinds of wrongfully harmful illocutions. Let us focus on the paradigmatic case of legal authority, with two examples made by Langton herself. The first, already encountered, case consists in a legislator uttering “Blacks are not permitted to vote” in an Apartheid regime (Langton Reference Langton1993, 302). The second consists in a husband saying “I divorce you, I divorce you, I divorce you” to his wife in a jurisdiction where Sharia law is in force and thus, allegedly,Footnote ²⁴ where saying so counts as unilateral divorce. The same is not true if the wife says the same to her husband: “Divorce of that kind is an act that is unspeakable for women” (Langton Reference Langton1993, 317). Now, what we normally regard as appropriate counter-measures in cases such as these is not to prevent people from speaking.Footnote ²⁵ We do not legislate that people are no longer allowed to say that Blacks should not be allowed to vote. We legislate, instead, that no such claim should ever count as the disenfranchisement of Black citizens (for instance, via constitutional constraints on legislation). Similarly, in the Islamic divorce case, the point is not to prevent married men from saying, “I divorce you, I divorce you, I divorce you” to their wives, but to not make that count as unilateral divorce. Once the relevant legislative change is in place, husbands can still yell “I divorce you, I divorce you, I divorce you” at their wives as much as they want; their wives, however, can perfectly well respond, “Fine, you had to get that out of your system. Now, for the real thing, talk to my lawyer.” In both cases, the relevant speakers have not been deprived of the power to pronounce certain locutions—they have, instead, suffered (justified) illocutionary disablement. The same reasoning applies, mutatis mutandis, to pornography: if pornographers have no claim to their epistemic authority, we must dismantle the felicity conditions that make their speech an exercise of authority. When this occurs, pornographic material may very well carry on portraying phantasies of women who say “no” but are thereby just engaging in foreplay; this, however, will no longer count as a standard-setting claim about what women are “really” like.

This can also be true for epistemic authorities. Suppose, say, that the APA enjoys epistemic authority without good reason; its view on mental disorders should therefore no longer be seen as standard-setting. This does not lift the strong presumption against censoring their manual—it merely implies that the manual must be deprived of its misplaced authoritative status. It should still be legally permitted to sell it and read it (including in schools and universities, under certain conditions and with the right disclaimers), but we might need to legislate against it being adopted as a textbook in schools, or as reference in official psychiatric protocols and diagnoses. None of this amounts to censorship: it merely means downgrading a text from epistemic authority to mere expression of a view.

Are things really that straightforward, though? A cluster of rejoinders might be raised at this point. One could object, first, that censorship may be justified, as a non-ideal strategy of sorts, if we have reasons to be skeptical about the prospects of the “ideal” strategy of illocutionary disablement. If the harm committed by pornographers is real, and we do not know for sure whether tackling felicity conditions will work, are we not justified in preventing them from doing harm in the most straightforward way, that is, by not letting them speak? The same, after all, is true in the Apartheid and divorce cases: if we do not know whether we will manage to implement legislative change, it seems that seeking to prevent Apartheid legislators and husbands from speaking is justified after all, at least for the time being. There is, however, one important disanalogy between the two considered examples and pornography, which actually makes the case for censoring porn harder. Pornographers, as we have seen, enjoy authority of an epistemic kind (if at all). When practical authorities—such as Apartheid legislators or husbands in the hypothetical example—are prevented from speaking altogether, they are indeed thereby prevented from exercising their authority in that specific occasion. The husband who is prevented from saying “I divorce you, I divorce you, I divorce you,” cannot thereby perform the illocution of unilateral divorce. Epistemic authorities, instead, make us believe certain things to be true and certain types of behavior to be appropriate. Therefore, it seems that whenever the issue of censorship becomes relevant, it is because hearers already believe certain things because of certain (past) speech acts—it is a matter of undoing harms, rather than straightforwardly preventing them from happening.Footnote ²⁶ What is more, the power of epistemic authorities is sticky: if hearers are under the spell of the epistemic authority of pornographers, they will not just cease to believe certain things the moment pornography is no longer around, like a snap of a wizard's fingers which can undo a spell.Footnote ²⁷ This is especially true for epistemic informal authority, which cannot be removed simply by withdrawing an official investiture—although it is doubtful whether that is ever the case, even for formal authorities. At most, one could claim that censorship might, in the course of time, break the spell under which pornographers hold their audiences (a point to which I shall return below)—not that the harm stops the very moment they are censored. Some of this is true, in a certain way, also for practical authorities. If I physically prevent a judge from applying an unjust law during by breaking into a courtroom, I may block a very specific incidence of unjust harm, but to the extent that the law remains in place, the threat of injustice for all its potential victims remains. In the case of informal epistemic authorities, however, we can hardly ever prevent even individual harms in that way. Thus, taking apart felicity conditions must be the ultimate goal.

Now, one could argue that preventing pornographers from speaking, at least for some time, might do just that—by no longer being exposed to porn, hearers might progressively cease to regard it as authoritative. The first thing that should be noted about this rejoinder is that it actually accepts one of the central claims of this paper: censorship is here suggested as a means to erode the wrongfully placed epistemic authority, rather than something in principle legitimate simply because we are no longer in the domain of free speech. This key point is grossly underappreciated in the debate. Yet, if the remark implies that censorship is something that we can just try out, hoping it might work, then it must still be rejected. If the speech-act argument against porn wishes to remain compatible with freedom of speech, as it does, we cannot recur to censorship casually: doing so will curtail the freedom of certain speakers to express views and opinions, hence the bar must be significantly higher than that. Things are different, of course, if we have compelling reasons to believe that it might be the only successful strategy. If this is actually the case, and if the harm of porn is significant enough, then we may indeed have reasons for recurring to censorship after all, despite the presumption against censorship. For this argument to fly, however, we must have strong evidence, first, that censorship will work; and, second, that dismantling felicity conditions alone, without preventing locutions, will be impossible or nearly so. For instance, confirmation biases might be so deeply enrooted that the mere availability of pornographic material showing women enjoying rape is a prohibitive obstacle to the erosion of the authority of porn. Thus, letting pornographers speak means letting them perform certain illocutions—under the current circumstances, their speech cannot but be taken as authoritative. If these two conditions obtain, we might say that censorship is, sadly, necessary for harm prevention.

There may be cases where the only way of depriving a text of its dangerous, unjustified authority is to remove it from circulation altogether, at least for a while; this was deemed to be the case, for instance, with respect to Mein Kampf in post-war Germany. This however, does not indicate that the presumption against censorship is lifted in cases of misplaced epistemic authority, and we can thus openly debate the opportunity of recurring to it. It just means that censorship may be viewed as an instrument of last resort to disable certain illocutions under some very exceptional circumstances, despite the strong presumption against it. This, however, would not be a new argument for censorship. Several—if not all—free-speech defenders already accept that, if censorship is the only way in which we can counter some very real and grave harm, we may have to recur to it in some exceptional cases. Langton has provided a powerful argument to show that pornography can indeed constitute harm; and such argument could very well constitute the premise of a complex, empirically informed justification for the (presumably temporary) censorship of porn in some extreme cases. This, however, is very different from the way in which the speech-act argument about porn is commonly understood to affect the very standing of censorship and the relationship between porn and freedom of speech.

Now, whilst the jury is by all means out on the matter whether censorship is necessary in this case, it may be worth pointing out that current empirical evidence in support of the epistemic authority of porn identifies the problem in its omnipresence and in the absence of easily accessible alternatives, rather than in its sheer availability (Maddy et al. Reference Maddy, Liz Kelly, Garner and Kanyeredzi2013; Doornwaard et al. Reference Doornwaard, den Boer, Vanwesenbeeck, van Nijnatten, ter Bogt and van den Eijnden2017). Thus, it seems to offer prima facie support for the regulation of access to porn (especially by certain salient groups, such as teenagers), possibly in much more extensive ways than is currently the case (more on this in the Conclusion), rather than for censorship outright. What if, however, conceptual reasons, with no need for empirical evidence, were sufficient to prove that measures short of censorship are inadequate in the case of porn? This is the question to which I now turn.

Is pornography special?

In this section, I respond to two attempts to establish that in the case of porn, tackling felicity conditions alone simply cannot work, for reasons that are conceptual rather than just empirical.

The first runs as follows. Dismantling the wrongful authority of pornographers without preventing them from speaking is impossible if women are not only subordinated, but indeed also silenced by porn. Some of Langton's choices of words can be seen as gesturing towards this view, and she is fairly often interpreted in this way by critics, although recent writing seem to clarify unequivocally that her view is much more careful (Reference Langton, Gardner, Green and Leiter2018a). Either way, the view is worth addressing. Langton recognizes the power of speech as a means of blocking wrongfully harmful illocutions (Reference Langton1993, 313–14; Reference Langton2016, Reference Langton, Gardner, Green and Leiter2018a). For this to be an option, however, it must be possible to fight speech with more speech. If women are silenced, however (rather than merely subordinated), how can they “talk back” (Butler Reference Butler1997) in the way that matters? “If pornography not only subordinates but silences women,” Langton remarks, “it is not easy to see how there can be any such fight […].Whether women can fight speech with more speech depends on whether, and to what extent, women can speak” (Reference Langton1993, 313–14). Notably, Langton does not linger on justifying this point much. Yet, this claim has been largely taken to imply that “the solution preferred by many liberals (and feminists) of countering the harms of pornography with more speech-protest, satire, education and public debate” (West Reference West and Zalta2004) is simply delusional if pornographers can silence, for women simply cannot engage in such fight.

I believe this conclusion to be unwarranted. Recall how the alleged authority of pornography is plausibly partial; as such, it is not immune from counterattacks. Women who are illocutionarily silenced by pornography suffer something close to Fricker's (Reference Fricker2007) testimonial injustice: they cannot count on succeeding in their attempted illocutions with the reliability and robustness which they deserve. They can expect to be often misunderstood, not taken seriously, etc. In Fricker's wording, they suffer an unfair deficit in credibility and reliability, rather than watertight illocutionary disablement; theirs is a significant impairment rather than an impossibility. This means that there is a window of opportunity to be explored and exploited. To have a sense of how this might unfold, consider a scene from the 1968 Planet of the Apes movie. The human astronaut who has crash-landed on the “planet of the apes” is captured and restrained; however, unlike other humans on the planet (who are mute and lead primitive existences) the astronaut is articulate, and makes an eloquent case for the equal rationality (and therefore moral standing) of humans and apes. The hardliners within the ape community dismiss his words, saying that he is simply “parroting” rational argumentation, without actually understanding what he is doing.Footnote ²⁸ Clearly, however, his plea leaves an impression on many apes in the audience, instils doubts in them, and ultimately wins him some precious allies. This is a very clear case of an attempt at illocutionary silencing proper which, although starting from a background position of authority, ultimately fails because its victim succeeds in eroding the relevant felicity conditions. What is more, if actors other than victims may also engage in this erosion attempts—for instance, in the case of pornography, men—they might stand an even better chance to succeed.

Of course, if speakers are not only subordinated, but also silenced (if imperfectly so), engaging in this kind of resistance will prove more difficult for them—and possibly for third parties, as well. It could be that countering silencing by tackling felicity conditions alone is prohibitively hard in many cases. Speech as an instrument against wrongful authoritative speech has, after all, a structural disadvantage in all cases, for undoing conversational norms is much harder than initiating them (McGowan Reference McGowan2009; but see Lepoutre Reference Lepoutre2019). My claim is merely that the mechanisms underpinning subordination and silencing are ultimately the same, and so are the means to fight both. What is more, tackling felicity conditions with means other than censorship need not—indeed, most likely will not—be a matter of isolated attempts. It will most likely entail regulating access to porn; making such access conditional to the reading of certain disclaimers or exposure to alternative sources; and injecting the public with healthy doses of alternative sources of information in a concerted manner.

Let me now turn to the second objection. According to it, porn is importantly unlike the cases discussed in the previous section due to the very informal character of its authority. In formal cases, we can simply invalidate the legal status of certain statements;Footnote ²⁹ things are different when a speaker enjoys authority through informal norms and mechanisms. This is not a mere issue of feasibility. When certain power equilibria are in place, changing laws is incredibly hard—as slavery abolitionists and suffragettes know all too well. The point is rather that, when the power is sustained by social norms, one has to erode those very social norms in order to take it apart, and this raises epistemic obstacles, not just sheer issues of power inequality. We very often simply do not know how to dismantle felicity conditions. And until we do, certain actors saying certain words, it seems, will unavoidably count as them performing certain illocutions for some hearers—the only way to stop them, or so the argument goes, is to prevent them from speaking altogether. Letting them speak means letting them perform certain (wrongful) speech acts.

To illustrate this point, let us return to the divorce example. Let us suppose that we have deprived the utterance “I divorce you, I divorce you, I divorce you” of its legal power. Let us also suppose, however, that social norms have not followed suit. A married man saying “I divorce you, I divorce you, I divorce you” to his wife still de facto counts as performing unilateral divorce in the community. Married men are still widely perceived as being entitled to initiate divorce unilaterally in this way in a way that is not available to women; de facto still do it; and women in the community, by and large, comply, by leaving the marital home without protesting (and possibly without advancing any claims) when these words are uttered to them. Now, is the authority of pornography not similarly sustained by informal social norms? And if it is, is it not true that pornographers cannot but be perceived as making authoritative statements as soon as they speak? And, finally, does this not imply that the only way to allow other “stories” on sex to have a chance is to remove the dominant “single story” from the shelf for a while?

Now, it is certainly true that, when the authority of harmful speakers is sustained by informal social norms, dismantling felicity conditions is particularly tricky. Again, however, this implies neither that it is impossible, nor that censorship will, instead, succeed. If anything, the example further illustrates that the capacity of some speakers to enjoy power over their hearers’ minds—even if the law says they should not—is what allows them to do harm with their words. Thus, if anything, it casts some doubt on the efficacy of legal means (of which censorship would be one) in these cases. We are thus back where we started: pornography could be special—such that censorship as measure of last resort, in a way that most free-speech advocates would also accept, is warranted; but the case still needs to be made, and it must be an empirical one.Footnote ³⁰