A Comprehensive Definition of Illocutionary Silencing

Abstract

A recurring concern within contemporary philosophy of language has been with the ways in which speakers can be illocutionarily silenced, i.e. hindered in their capacity to do things with words. Moving beyond the traditional conception of silencing as uptake failure, Mary Kate McGowan has recently claimed that silencing may also involve other forms of recognition failure. In this paper I first offer a supportive elaboration of McGowan’s claims by developing a social account of speech act performance, according to which the success of an illocutionary act is not only a function of the intentions of and the conventions deployed by the speaker, but partly depends on how the act is recognized or taken up by the hearer. I then provide a comprehensive definition of illocutionary silencing and spell out what it means for it to occur in a systematic manner.

1 Introduction

One may give various narratives about speech. The first goes like this: speech is a matter of moving one’s mouth and making intelligible sounds—or, of moving one’s fingers and typing significant strings of letters on a computer keyboard. On such a narrative, a person’s power to speak may be narrowed down only by such things as intimidation or threat of harm, or by material obstacles like gagging, knocking out, or withholding access to computers. If speech is nothing but saying things, preventing someone from uttering words is the one and only way to silence them. However, as Austin (1975 [1962]) famously pointed out, a more faithful narrative about speech and its nature would shape it as a sort of action. Speech, this alternate story suggests, is not only a matter of saying things, but also and foremost of doing things with words. Conceiving of speech along these lines offers a different way to construe the notion of silencing. One can be free to talk and yet be deprived of, or otherwise impaired in, one’s capacity to perform certain speech acts. By drawing on Austin’s notion of illocution as the performative aspect of language use, such a peculiar sort of silencing has been labeled ‘illocutionary silencing’.

The Austinian approach to silencing was introduced by Jennifer Hornsby and Rae Langton (henceforth, H&L) in the early ’90s to defend the philosophical plausibility of Catharine MacKinnon’s silencing thesis against pornography.¹ It is well known that MacKinnon objects to pornography on the ground that it silences women. By extending Austin’s insight that words do things, H&L have first argued that MacKinnon’s use of the term ‘silencing’ is primarily meant to capture a failure to act. It is not that pornography prevents women from producing locutions; it is that it undermines their capacity to do things with those locutions. Specifically, pornography may make it nearly impossible for women to refuse unwanted sex. Pornographic materials may help create an uncomprehending communicative climate that interferes with men’s ability for uptake, i.e. their ability to recognize the illocutionary intention behind a woman’s “No” to sex. Since in Austin’s framework the hearer’s uptake is necessary for illocuting, a woman’s “No”—not understood for what it was (a refusal) or misunderstood for something else (a consent)—ends up misfiring. The thesis that pornography silences, H&L contend, is philosophically plausible.

Here, I will not deal with MacKinnon’s anti-pornography claim but rather focus on silencing² per se. While H&L have framed it in terms of uptake failure, Mary Kate McGowan has later maintained that silencing may also involve other forms of recognition failure.³ A speaker may be sincerely speaking, but her sincerity may go unrecognized; or she may be authoritatively speaking, and yet be taken to lack authority. When this is the case, one may suffer silencing, albeit uptake has been secured. This paper has a twofold aim. First, I offer a supportive elaboration of McGowan’s claims by arguing that, if we take the social character of illocution seriously (if we assign both the speaker and hearer an active role in illocutionary performance), then it is altogether possible that a hearer’s failure to recognize some pragmatic constituents of a speaker’s illocutionary act throws that act off the rails. Second, I provide a rigorous definition of silencing—something remarkably missing in current philosophizing over silencing. Although the silencing literature is by now rather extensive,⁴ scholars have been mainly concerned with advocating for or against the claim that pornography silences, leaving the question of what precisely silencing amounts to pretty much untouched. I draw up a broad enough definition to cover different kinds of recognition failure, while ruling out linguistic breakdowns which would seem intuitively wrong to count as silencing. It is generally agreed that silencing must satisfy a systematicity condition, but what ‘systematicity’ means in this context has been far from clear. I propose to shape the notion of systematicity by analogy with what Grice (1975) called ‘non-detachability’, thereby departing from the characterization in terms of ‘widespread beliefs’ that has been assumed in the debate so far.

I begin by offering some background in speech act theory, thus preparing the terrain for a theoretically grounded analysis of silencing, and proceed by putting forth a social, as opposed to an individualistic, account of speech act performance (Sect. 2). I then move on to disentangle the main pragmatic constituents of illocution and the types of silencing arising from the hearer’s failure to recognize them. In doing so, I will provide examples other than the act of sexual refusal to show that the notion of silencing can be fruitfully applied beyond the sphere of sexual negotiation (Sect. 3). With all this in place, I put forth my definition of silencing and conclude by discussing the meaning that ‘systematicity’ has in this context (Sect. 4).

2 Theoretical Terrain

2.1 Locution, Illocution, Perlocution

The trend in the philosophy of language referred to as ‘speech act theory’ has arisen out of a central insight—namely, that saying something is doing something. But ‘doing something’ is quite a vague expression. In order to make it more specific, Austin distinguished between three broad kinds of speech acts, to wit, locutionary, illocutionary, and perlocutionary acts.

In Austin’s parlance, ‘locution’ captures what saying consists in—i.e. the utterance of a sentence with a certain sense and reference. To perform a locutionary act is eo ipso to perform an illocutionary act, at least in standard cases. Illocutionary acts correspond to the actions the speaker performs in uttering certain words, or differently put, to the peculiar force of the locution in the context of utterance. Consider, as way of an example, an utterance of

(1) Shoot him!⁵

It is pretty clear that the speaker is here saying something (i.e. performing a locutionary act) as well as doing something—i.e. ordering the hearer to shoot a certain (salient) man. The speaker, that is, is also performing the illocutionary act of ordering. Perlocutionary acts complete the picture of what we can do with words. They refer to the relation between the things we say and the bundle of effects those things produce on our audience’s thoughts, feelings, and behaviors. Imagine that, by uttering (1), the speaker causes the hearer to actually fire a hail of bullets or makes the target man’s heart start pounding. Consequential effects like these inhabit the realm of perlocution.

Locutionary, illocutionary, and perlocutionary acts are not fully-fledged speech acts, but abstractions that in fact co-exist within a single speech act performance. Yet, they may fail independently of one another. For our present purposes, I shall concentrate on illocution and gloss over the types of failures locutionary and perlocutionary acts are characteristically susceptible to. Austin (1975 [1962], p. 14f) introduced three sets of rules whose violation makes an illocutionary act infelicitous. An attempt to illocute misfires if the invoked conventional procedure does not exist or an existing procedure is invoked by inappropriate persons or in improper circumstances (A rules). Imagine that one attendee to Rafael and Beth’s wedding ceremony, completely drunk, interrupts the service to pronounce the couple husband and wife. Since the drunk attendee is the wrong person to wed the couple, his utterance results in a void act. Further fatal infelicities involve the execution of the procedure, which must be carried out correctly and completely (B rules). An act of marrying, for instance, is correctly performed only if the parties utter certain formulae. Finally, an illocutionary act is an abuse of the procedure—it is defective albeit not a complete failure—when the speaker’s performance is insincere (e.g. thanking with no gratitude) or her subsequent behaviors are inconsistent with it (e.g. advising someone to do φ and then reproaching them for doing it) (Г rules).

Austin’s felicity rules have later been recast in terms of success conditions (Searle 1969; Searle and Vanderveken 1985). Austin’s A rule on the appropriateness of participants and circumstances flows into Searle’s ‘preparatory conditions’—i.e. states of affairs that must obtain in the world of utterance for an illocution to be successful. Consider again the act of marrying. A marriage ceremony can be officiated only by a speaker endowed with the right sort of authority. In Austin’s terms, marrying complies with an A rule prescribing that the speaker is an authorized minister; in Searle’s terms, marrying comes with a preparatory condition requiring speaker authority. Similar remarks apply to Austin’s Γ rule concerning insincerities or dissimulations, which maps directly onto Searle’s ‘sincerity conditions’–i.e. the psychological state(s) the speaker must have for her performance to be a sincere one.

2.2 Illocuting as a Two-Sided Process

It is worth adding at this point that Austin’s felicity rules, as well as Searle’s success conditions, may be broadened to cover uptake—which Austin took to involve the audience’s recognition of the meaning and force of the locution, and Searle reframes in a Gricean-inspired way as the audience’s recognition of the speaker’s illocutionary intention.⁶

If I am trying to tell someone something, then (assuming certain conditions are satisfied) as soon as he recognizes that I am trying to tell him something and exactly what I am trying to tell him, I have succeeded in telling it to him. Furthermore, unless he recognizes that I am trying to tell him something and what I am trying to tell him, I do not fully succeed in telling it to him (Searle 1969, p. 47).

Even though both Austin and Searle qualify uptake as an illocutionary effect,⁷ its achievement is deemed to be necessary for an illocution to be fully successful. Taking this seriously, illocuting comes across as an essentially two-sided process, involving a speaker and an audience. One cannot perform an illocution in a fully successful way without one’s audience actively grasping the kind of illocution that it is and, I claim below, its core components.⁸ Note in passing that a social approach to illocution does not preclude the possibility of inner speech acts. Suppose that, after a wicked hangover, I promise myself I will never drink again. Do I acquire a genuine commitment not to drink ever again? This is a controversial issue, but if the answer is yes, then promises to oneself are genuine illocutions. I think, however, that they would only apparently be ‘solo’ illocutions. Their performance would still be a two-sided process, involving a speaker and a hearer who happen to be the same person. Soliloquy, in such cases, would take on the contours of an ‘interior poliloquy’.⁹

I work within a conception of language use according to which our illocutionary acts are such that they reshape the normative context they occur in by (un)assigning new deontic roles to certain relevant parties.¹⁰ A successful promise commits the speaker to a future course of action, an order imputes an obligation upon the addressee, and a fully-fledged assertion carries a burden of proof that a speaker must be able to bear. When no official ratification is involved (as happens with law enactments, for instance), the coming into existence of a new normative status partially depends on whether the audience takes the speaker’s utterance in a certain way and recognizes that certain statuses have been thereby engendered. If you promise me that you will quit smoking while I am sleeping or I am wearing earplugs which isolate me completely, then it is not at all clear that a commitment for you to quit smoking has therewith been created.¹¹ I interpret the role of uptake in the light of the normatively transformative power of illocution. The normative statuses introduced by illocutionary acts are relational in nature: if you promise to do something, you acquire a commitment to the hearer; therefore, insofar as the hearer does not realize that you were promising, one may doubt that your utterance introduced any commitment at all. In what follows, I broaden such an insight and argue that several different elements of an illocutionary performance must be recognized by a hearer for that performance to be fully successful. Suppose you promise me that you will quit smoking and I perfectly understand what you are saying and trying to do with words, but believe that you are lying to me. After all, you keep promising that you will quit smoking and break the promise every time. Suppose, however, that this time you have a firm intent not to smoke a single cigarette anymore. Would your promise, mistakenly taken to be an insincere one, commit you to make good on it? The answer, I think, is yes—but it is intuitively clear (and theoretically sound, as I will explain later) that my failure to recognize your sincerity somehow affects your performance by making it less than fully successful.

This will suffice to lay out the terrain. In Sect. 4, I will define silencing on the background of the social account of illocution just sketched. But before getting to that, let us have a closer look at the different features of a speech act performance that can go unrecognized by an audience. Such an analysis will allow us to identify various types of silencing, which I will later group under a unique, inclusive definition.

3 Distinguishing Types of Silencing

The class of illocutionary acts is highly heterogeneous. It comprises ordinary acts such as telling or stating, which at first glance are no more than acts of saying, as well as institutional acts such as bequeathing or vetoing, whose performative character is far more visible. Silencing typically involves ordinary illocutions, impairing speakers in their capacity to perform acts that they would otherwise have been able to easily perform in their everyday interactions. In what follows, I focus solely on ordinary (i.e. non-institutional) speech acts, leaving aside the question about whether institutional acts have a different pragmatic texture than ordinary ones.¹²

The pragmatic constituents of an ordinary act performance include at least (i) the speaker’s illocutionary intention, (ii) the speaker’s sincerity, (iii) the speaker’s seriousness, and depending on the kind of act performed, (iv) the speaker’s authority. The hearer’s recognition of each such constituent, as we will see shortly, bears on illocutionary success, and her systematic failure to recognize this or that constituent originates different types of silencing.

3.1 Speaker Intention

It is traditionally accepted that, for a speaker to successfully perform an ordinary illocution, they must intend to perform it. One of the clearest expressions of this is found in Daniel Vanderveken.

Because illocutionary acts are intrinsically intentional, no speaker can perform an illocutionary act in a context of utterance unless he makes an attempt to perform that act in that context (Vanderveken 2002, p. 146).

Note that ‘illocutionary intentions’ are distinct from ‘perlocutionary intentions’. An illocutionary intention is an intention to perform an illocution of a given kind; a perlocutionary intention is an intention to affect the hearer in some way—to get her to believe that p or to do φ. Imagine that, while waiting at the tram stop, Julia says to the only guy waiting with her,

(2) Very nice weather today, isn’t it?

Julia successfully remarks that the weather is very nice only if she has the (illocutionary) intention to do so and such an intention is recognized by her interlocutor. It is very likely the case that, in a situation like this, Julia has no (perlocutionary) intention to make her interlocutor believe that the weather is nice—or, for that matter, to activate that belief in him. She is just ‘making conversation’, as Alston (2000, p. 49) puts it, and trying to avoid the social awkwardness of silence. As one can see, not only are illocutionary and perlocutionary intentions conceptually distinct, but one can intend to perform a certain act (and succeed in so doing) without having the perlocutionary intention that is standardly associated with it. Thus, the success of an ordinary act (partly) depends on the speaker having the relevant illocutionary intention and the audience recognizing that intention.

Silencing, as originally conceived by H&L, interferes precisely with the audience’s recognition of the speaker’s illocutionary intention. It captures cases where one’s illocutionary intentions systematically fail to be recognized, despite one’s best efforts. The hearer may have no clue about what kind of act the speaker is attempting to accomplish, but more often they will misunderstand its nature by attributing to the speaker an intention that she does not have. The paradigmatic case of silencing-as-uptake-failure concerns, as I mentioned at the outset, women’s sexual refusal.

A woman says “No” to a man, when she is trying to refuse sex; she uses the right locution for an act of refusal, but somehow her speech act goes wrong […]. She says “No”, intends to refuse, but there is no uptake in her hearer. She is therefore not fully successful in refusing (Hornsby & Langton 1998, p. 27).

Due to certain sexist myths about women’s behaviors in sexual settings, some men may fail to grasp a woman’s intention to refuse sex and indeed interpret her “No” as if it were a consent.

The story of Linda Marchiano’s Ordeal can also be seen to involve this type of silencing. After starring in the famous porn movie Deep Throat, Marchiano (performing as ‘Linda Lovelace’) wrote a book titled Ordeal to tell of how she was beaten, raped, drugged, hypnotized, and intimidated into performing her starring role. Ordeal is a forceful denunciation of the porn industry, an act of protest against the exploitation of women that underlies certain porn productions. Yet, it was advertised and sold in R-rated catalogs. Why? Part of the answer is that it was taken to be pornography, at least by some. The case is a striking example of silencing-as-uptake-failure, where not only did people fail to recognize Marchiano’s illocutionary intentions; they also distorted those intentions in a way that further exacerbated the precariousness of her position.¹³

3.2 Speaker Sincerity

It is one thing to intend to perform a certain act. It is quite another thing to intend to comply with the commitments created by one’s performance or to actually have the psychological states that performance expresses. This can be easily seen in the case of promises. One may utter a sentence intending to promise but lack any intention to do what one is promising to do. Promises, and (ordinary) illocutions more broadly, may be performed insincerely. Consequently, besides H&L silencing, there may be a further type of silencing occurring when the addressee, while recognizing the speaker’s illocutionary intention, mistakenly believes that she is acting insincerely. The locus of the failure is not the primary interpretation of the type of act the speaker tries to carry out, but whether it is sincerely performed or not. McGowan (2014) argues that women’s sexual refusals may be silenced in this way as well: while in H&L’s ‘uptake failure account’, the man fails to take the woman’s “No” as a refusal altogether; in McGowan’s ‘insincerity account’, he takes it as a refusal, but thinks that she is not sincerely refusing. If the man did not see the woman’s locution as a refusal, he could not have seen it as an insincere one. The former is necessary for the latter.

The ineffectiveness of the woman’s “No” might also be explained in a slightly different way, still compatible with the insincerity account. Suppose the man recognizes that she intends to sincerely refuse but thinks that refusing is not what her ‘deep self’ wants. Taking himself to be acting according to her true feelings, he dismisses her refusal and goes ahead with sexual advances. Although there is a sense in which the man here recognizes the woman’s sincerity, I regard this as a variation on sincerity silencing. After all, he only recognizes her surface sincerity. He understands that she is not faking it but takes what she presumes to be the case to drastically diverge from what is actually the case. That is, he takes her not to be truly sincere: if she searched her soul, she would agree that she does not want him to stop.¹⁴

Several authors have doubted that ‘sincerity silencing’ is a genuine form of illocutionary silencing.¹⁵ Speaker sincerity is not necessary for illocuting (it is always possible to perform an insincere speech act by expressing a psychological state one does not have); nor (a fortiori) is its recognition. When the speaker’s sincerity goes unrecognized, her illocutionary act may still fully succeed: the failure lies at the perlocutionary level. I think, however, that more is to be said here. Illocutionary acts, says Austin, “invite by convention a response or sequel” (1975 [1962]), p. 117). For example, a question invites an answer and an order invites obedience. The inviting of such responses is connected to the performance of those illocutions “by convention”. In the deontic framework for speech acts adopted here, this means that any illocutionary act type invites a particular kind of response in virtue of its normative effects. It is because an order imputes an obligation upon the addressee that it invites obedience, and it is because a question gives the addressee a petitionary reason to provide the information the speaker has asked for that questions invite answers. Differently put, the normative upshot of an illocution standardly translates into the inviting of a specific kind of response or sequel (which, if accorded or implemented, will satisfy the act’s perlocutionary object). I argue that, in sincerity silencing cases, the speaker’s act is rendered unsuited to invite the appropriate response on the part of the hearer. In the sexual refusal scenario, the woman’s “No”, mistakenly taken as insincere by the man, will succeed in imputing an obligation upon him to stop, but such an obligation will not translate into an invitation for him to actually do so. From his point of view, she wants to have sex but is feigning unwillingness whether outright or on a deeper level. He will thus think his conduct is in line with her desires, and her “No” will not dispose him to back off. Once we emphasize that illocutionary acts conventionally invite a response, something Austin acknowledged but that has gone largely overlooked in the subsequent literature, sincerity silencing plausibly turns out to be a species of illocutionary silencing.

Before moving on, note that this type of silencing is likely to be involved with other sorts of speech acts too. Examples might be that in a court of law the testimony of a black witness is misjudged and perceived as insincere because of racial prejudices and biases, or that the police do not believe a rape accuser because they falsely think that women tend to lie about rape.¹⁶

3.3 Speaker Seriousness

Speech act theorists generally consider seriousness as a precondition for illocution. An utterance is well suited to constitute an illocutionary act only if it is seriously spoken—only if the speaker is not joking, play-acting, rehearsing a speech, and so on.¹⁷ The failure to recognize the speaker’s seriousness may give rise to another type of silencing. Elsewhere, I have called it ‘seriousness silencing’.¹⁸ While the literature has tended to flatten it out into silencing-as-uptake-failure, I claim that it is a genuinely distinct type of silencing. To begin seeing this, consider a much-discussed case from Davidson (1984, p. 269). A theatrical production, the actor on the stage is shouting for fire as part of the play.

(3) Fire! I mean it! Look at the smoke,

says the actor, following the script. In the meantime, a real fire breaks out. The actor tries to warn the audience by uttering (3), but he does so in vain: the audience remains in their seats.

Langton (1993, p. 321; 1998, p. 275) treats this case on a par with silencing-as-uptake-failure. The woman who fails to secure uptake for her “No” would be in the plight of Davidson’s actor: her illocutionary intention to refuse does not get through to the man, just as the actor’s intention to warn does not reach out to the audience. Unlike Langton, I think the two accounts should be kept separate. In the ‘uptake failure account’, the hearer recognizes that the speaker is illocuting—albeit he does not grasp the nature of the act she is trying to accomplish. In the ‘play-acting account’, the hearer takes the speaker to be performing a non-serious act, that is, not to be illocuting. To word it differently, in the ‘uptake failure account’, the hearer fails to recognize that the speaker intends to perform a certain act (e.g. that she intends to sexually refuse); in the ‘play-acting account’, he fails even to recognize that she intends to illocute.

Notice, moreover, that seriousness silencing does not collapse onto sincerity silencing. Play-acting clearly has something in common with insincerity: in both cases, the speaker is somehow pretending. It is crucial, however, to distinguish two different senses of ‘pretend’. In one sense, to pretend₁ is to engage in a speech act performance which presupposes one has a psychological state one actually lacks; in the second sense, to pretend₂ is to engage in the performance of a speech act one is not really performing. If I lie to you, saying that I haven’t drunk any alcohol since last weekend when I just had a glass of wine before you came along, I am pretending₁. My performance is successful though defective: a lie is an assertion after all. It is quite different if I do the same thing—telling that I haven’t drunk any alcohol for the last few days when I just did—while I am acting a part in a play. In pronouncing my line (“I haven’t drunk any alcohol since last weekend”), I would make as if to assert; I would be engaging in a pseudo-performance which constitutes pretending₂ to do such and such.¹⁹ Seriousness silencing, then, does not conflate with sincerity silencing: sincerity silencing has to do with the hearer mistakenly believing that the speaker is pretending₁; seriousness silencing involves a false belief that she is pretending₂.

3.4 Speaker Authority

Certain illocutionary acts require that the speaker have some sort of authority. The floor manager at a machine factory can order a subordinate to carry out certain tasks, but a subordinate cannot order her, say, to reduce his working time or to move his lunch hours. If he did so, his purported orders would misfire, for he would lack the requisite authority. Importantly, they would misfire even if the manager grasped that he was seriously trying to order. McGowan (2009) identifies a peculiar type of silencing brought about by the hearer’s failure to recognize the speaker’s authority over a certain domain. While she introduces it to account for yet another way that women’s sexual refusals can go awry, I take the notion of ‘authority silencing’ to be useful to make sense of several impediments to illocuting. Before considering some of them, note that authority can be conceived of as something practical (i.e. a right to tell people, within certain limits, how to act) as well as something epistemic. While a floor manager has authority over the workers on her floor, there are also cases where speakers have “‘mastery’ of a subject matter, rather than of a human subject”, as Langton (2015, p. 18) nicely puts it. In its paradigmatic form, this is the authority of the expert, who can vouch for the reliability of some pieces of information. It is an authority about what to believe (as opposed to how to act) and is a matter of competence, or perceived competence.²⁰

Epistemic authority is a felicity condition for expert speech acts, which span from expert assertions about particular subject matters to advice within specific domains of expertise. I think we see epistemic authority silencing frequently when women speak, or try to speak, as experts in male-dominated fields. If someone is an aerospace engineer and makes a claim about the efficiency of a new spacecraft technology, their claim has an intrinsic weight. Though challengeable, it is credible in a way that a non-expert’s claims about spacecraft technologies are not. Often, in spite of being competent and thus satisfying the requirements for counting as an expert in a given area, a woman finds that her utterances do not count as expert speech acts. Her expert status is not recognized and hence fails to give hearers any special reason to trust what she claims. Moreover, male experts sometimes get to overrule what a female expert has claimed merely in virtue of their expertise, as they would do with a lay speaker.²¹

Let me stress that, though one must have authority to perform an authoritative illocution, the addressee’s recognition of that authority matters too. McGowan (2009, p. 492) argues for the significance of authority recognition by resorting to common intuitions: it is intuitive that an (otherwise fully successful) order will be defective or non-ideal if the addressee fails to recognize that the speaker has the requisite authority. I think there are reasons in support of this claim that go beyond intuitions. Recall that Austin and Searle interpret uptake in quite different ways—as the hearer’s recognition of the (meaning and) force of the locution, and as the hearer’s recognition of the speaker’s illocutionary intention, respectively. These two interpretations, I suggest, may be taken to capture two different types of uptake, both of which impact on illocutionary success. As hinted above, an utterance can have a certain force only insofar as it imputes certain normative statuses upon relevant people. For example, an utterance can have the force of an order only insofar as it imputes an obligation upon the hearer to do what she was ordered to do. If this is right, then for a hearer to recognize the force of an utterance, she must recognize how it changes the normative situation—what kinds of rights, duties, entitlements, and the like, it imputes on people. When such a recognition does not occur, ‘Austinian uptake’ fails, and the act is not fully successfully performed. In ideal circumstances, the two kinds of uptake go hand in hand: the hearer understands that the speaker intends to order and that the utterance counts as an order. But in non-ideal circumstances, they may break apart: the hearer may understand that the speaker intends to order but fail to realize that her utterance counts as an order. This is what happens in authority silencing cases: the hearer grasps the speaker’s intention, but since he mistakenly believes that she lacks the authority to perform the act in question, he fails to recognize that her utterance counts as that very act. He fails to see how it changes the normative situation, with the result that Austinian uptake is not achieved and the speaker is not fully successful in illocuting.

4 Defining Silencing

4.1 From Uptake to Recognition

I have discussed four types of silencing and argued that each type is constituted by some sort of recognition failure on the hearer’s part. More precisely, silencing may involve a failure to recognize the speaker’s illocutionary intention, her sincerity, seriousness and/or authority over a particular domain—which are all key constituents of ordinary illocutions. I am now in a position to draw up a definition of illocutionary silencing that covers them all, and possibly leaves room for some other ways that illocutionary acts can be silenced. Since the constituents of an illocutionary performance also determine under what conditions that performance is fully successful, I frame the scope of the recognition failure that is involved in silencing in terms of success conditions.

Illocutionary Silencing: A speaker S putting forth a speech act A addressed to a hearer H is illocutionarily silenced iff (i) H fails to recognize the obtaining of some conditions for A’s success; (ii) S’s attempt at A-ing meets the conditions that H fails to recognize; (iii) normal input and output conditions are met; (iv) the recognition failure on H’s part is systematic.

As we saw, H&L introduce the notion of silencing by representing it as a form of uptake failure. Our discussion made it clear, however, that the notion should be given a broader characterization that moves from uptake to recognition failure. It is not only the speaker’s illocutionary intention that can go unrecognized, but also other important features of her speech act performance. When this happens (and provided certain further criteria are obtained), the speaker is silenced, even if she does achieve her audience’s uptake. Note that the audience can fail to recognize more than one feature of the act the speaker attempts to accomplish; this means that there may be complex instances of silencing where multiple types co-occur.

My definition entails that speakers can be silenced even when their act’s flaws are not uniquely traceable to recognition failures. Condition (ii) does not require that the speaker’s act fulfill all conditions for its success, but that it fulfill (at least) those conditions that the hearer fails to recognize. Suppose that Lois, a politician with a degree in nuclear physics, asserts that nuclear energy is dangerous and should be abandoned, but that she does so just for political purposes (she thinks it would bring in votes)—the truth being that she firmly believes the benefits of nuclear power outweigh its risks. Suppose also that Jim, her interlocutor, erroneously thinks that Lois lacks the requisite epistemic authority to assess the pros and cons of nuclear power and views her assertion as unwarranted. I take it that Lois may be illocutionarily silenced here, although her assertion’s defectiveness is partly due to authority recognition failure and partly to her insincerity.²²

The definition also rules out one-off idiosyncratic cases of uptake failure. Imagine that I am trying to tell Andrea that it’s his turn to cook dinner tonight, but Andrea does not hear because he is wearing headphones (as he usually does when he is writing). My attempt to tell him that it’s his turn to cook does not achieve any uptake and thus fails to be fully successful. Despite this, Andrea’s recognition failure does not and should not count as an instance of silencing. It would be stretching the notion of silencing a bit too far to suggest that one is silenced any time one’s interlocutor has impediments to hearing, does not share a common language with the speaker, or gets distracted by something happening while the other is speaking. (Suppose Andrea fails to understand what I am saying because he is distracted by the TV news showing footage of a bombing attack.) Cases like these are ruled out by condition (iii): one is silenced only if the hearer fails to recognize some aspects of one’s speech act performance despite the normal input and output conditions for communication obtained. The recognition failure, that is, must occur despite the background requirements for intelligible speaking and correct understanding are met.²³ Note that these also include linguistic competence. On the speaker’s part, this means that she must know what the conventions to perform the speech act in question are and employ them to perform it. Speakers cannot express themselves in any way they please and then complain about their audiences’ incapacity to understand them. A speaker who utters the sound ‘grumph’ to refuse, for instance, cannot complain that she has been silenced when the hearer fails to give her act the right uptake.

4.2 The Systematicity Condition

According to my definition, not every instance of recognition failure constitutes silencing. It is silencing only if it occurs in a systematic way. This is generally agreed in the literature, but what ‘systematicity’ means in this context is far from clear. Here is my proposal.

Systematicity Condition: an instance of recognition failure meets the condition iff it is attached to the illocutionary act S attempts to perform—and not to the linguistic form of the utterance S uses to perform it.

Consider again the sexual refusal case. The woman cannot (fully successfully) refuse whatever words she utters—“No”, “Don’t touch me”, “I don’t want to have sex with you”, and so on and so forth. If her male interlocutor is sure enough that, for example, women always desire sex but do not want to appear too forward, he will most likely fail to recognize her refusal as a sincere one no matter what she says. The recognition failure on his part will satisfy the systematicity condition: it will depend on the illocutionary act she attempts to accomplish in that context, and not on the particular word choice she makes.

The systematicity condition is here shaped along the lines of what Grice called ‘non-detachability’. Grice argued that conversational implicatures are non-detachable—they are attached to what is said rather than the way it is said, so that generally “it will not be possible to find another way of saying the same thing, which simply lacks the implicature in question” (Grice 1975, p. 58). If my partner requests that I walk the dog tonight and I answer, “I’m exhausted”, the implicated rejection would be maintained however the utterance had been reworded. “I’m really tired”, “I had a long day”—all carry the implicature that I’m not walking the dog tonight. While non-detachability appeals to a distinction between semantic meaning and linguistic form, systematicity rests on a distinction between illocution and linguistic form, so that it is generally not possible to find another way of performing the same illocutionary act, which easily avoids silencing. I follow suit with the Searlian tradition in conceiving of illocution as a compound of force and content. When an illocutionary act is silenced, it is typically the case that both its force and content are relevant to the hearer’s recognition failure. We saw that a woman who tries to refuse sex may be subject to various types of silencing. But a woman who tries to refuse, say, a job offer or a medical treatment may not meet any obstacle along the path to illocutionary success. It follows that the recognition failure underlying silencing is not tied to the act’s force alone. It is not tied to the sole content either. Suppose that a woman asserts that football referee X misjudged an offside position. Her male interlocutors, convinced that a woman would most likely not know the offside rule, might take her to be guessing that the referee was mistaken, or expressing confusion about whether the referee was mistaken, or else asking for their confirmation that the referee was mistaken. Had she made a guess, expressed confusion or asked for confirmation rather than asserting, her performance would have gone smoothly. Cases of this sort show that the recognition failure underlying silencing does not merely depend on the act’s content, but is attached to the illocutionary act as a compound of force and content.

This characterization of the systematicity condition predicts that silencing is quite hard to overcome or escape.²⁴ It is not enough to rephrase the sentence uttered—some extra effort is required to remedy the hearer’s recognition failure (if it can be remedied at all). Davidson’s actor might need to jump down into the audience or to turn on the lights in the theatre to signal that the conventions of fiction are suspended. And the woman in the sexual refusal scenario might need to physically resist to press home the message. Actions of this sort are no guarantee of success, though. The actor’s moves could still be interpreted as part of the play and women’s struggles against unwanted sex could be read, through the filter of rape culture, as implicit invitations for more. It is important to stress here that the possibility that the woman’s intentions eventually get through to the man confirms, rather than disproves, that she is illocutionarily silenced. Silencing erodes speakers’ capacity to do things with words—more precisely, with those words conventionally suited for the performance of the act in question.²⁵ The fact that a woman can resist does not mean that she can perform the illocutionary act of sexual refusal: resisting is not a way of performing that act and is not illocutionary in nature—albeit it may represent an extreme tool to make one’s intentions manifest. We can change the actual, material situation by fighting, by screaming for help, by running away. But these are non-normative attempts to stop something from happening or continuing. As such, they are profoundly different from illocutions of refusal, which are normative in nature.²⁶

The systematicity condition has traditionally been taken to require that the recognition failure on the hearer’s part be brought about by widespread beliefs, and therefore be likely to be made by others under similar circumstances.²⁷ My proposal departs from this characterization. I do not deny that many or even most instances of silencing are caused by widespread beliefs (typically, about the speaker’s social identity), but I think that this is not necessarily the case. Imagine that Oliver is trying to sincerely tell Arnold that he bought a big fancy yacht. Imagine moreover that Arnold has a bizarre belief that people with blue eyes often make up stories just to boast about themselves, and that Oliver happens to have blue eyes. As a consequence, Arnold does not believe what Oliver tells him. I tend to read this as a case of sincerity silencing, even though it is a peculiar fact about Arnold (and perhaps, about a few others) that he believes that people with blue eyes are compulsive liars. Arnold’s recognition failure is hard to escape on the spot: Oliver cannot easily rephrase his original utterance (“I’ve bought a Majesty 175”) in such a way that Arnold recognizes his sincerity. He could resort to further means to prove his sincerity, e.g. show Arnold some photos of his yacht or walk him to his yacht’s slip, but insofar as we look at Oliver’s capacity to felicitously tell Arnold that he has bought a luxury yacht, that capacity seems to have been curtailed. Note that Arnold’s failure does not derive from his incompetence in the language spoken by the speaker or from the fact that he got distracted while the speaker was speaking. It instead derives from a belief that Arnold strongly holds, and which unfairly interferes with Oliver’s ability to perform a fully successful act of telling. I do not see any reason why such an interference should not be counted as an instance of silencing. After all, Oliver’s illocutionary agency is unduly restricted—albeit only in the localized context in which his conversation with Arnold takes place. The recent literature has been nearly exclusively interested in cases of silencing owing to widespread negative identity prejudice. While I acknowledge the importance of focusing on such highly problematic cases, where silencing does not only wrong the silenced speaker but also contributes to the overall disadvantage of the social group the speaker belongs to, I advocate a notion of silencing which is theoretically independent from any unjust social structure. To silence an illocution is to unfairly prevent a speaker from fully successfully performing it. Although one’s membership in a relatively disempowered group renders one particularly vulnerable to silencing, a speaker’s illocutionary agency may be unfairly narrowed down for reasons that have nothing to do with her social identity. In a perfectly just social world, isolated episodes of silencing could still take place.

It is finally to be noted that my way of characterizing the systematicity condition renders the challenge of the ‘drowning case’, recently raised by McGowan et al. (2016), less insidious. The case goes like this: imagine that Sally is drowning and Peter, who is walking by, tries to save her. Sally says “No” intending to refuse his assistance. Peter hears what she says but believes that drowning people want to be saved, and thus interprets her utterance as an expression of denial (‘This is just too awful to be happening to me!’) rather than a refusal. This interpretive mistake is brought about by widespread beliefs about what people generally desire and how they behave when their life is at risk. Moreover, and relatedly, others would likely make the same mistake under similar circumstances. This leads to a dilemma: either the drowning case does not meet the systematicity condition, but it seems that it does (at least insofar as one adopts the traditional view); or it counts as an instance of silencing-as-uptake-failure, but “it just seems plain wrong to say that Sally is silenced” (McGowan et al. 2016, p. 79). The first thing to point out is that, if one decouples silencing from structural injustice, as I do, the claim that Sally is silenced—i.e. that her capacity to refuse in that particular context is restricted—appears less intuitively wrong. Other than that, though, my characterization of the systematicity condition makes McGowan et al.’s formulation of the dilemma untenable: the drowning case does not involve any systematic uptake failure. While the woman in the sexual refusal scenario cannot felicitously refuse whatever words she utters (e.g. “No”, “Don’t touch me”, “I don’t want to have sex with you”), Sally can refuse by fully manifesting her intention. It is plausible that, in a ‘drowning context’, an utterance of “No” would be interpreted as an expression of denial, but a full sentence like “No! I wanna die” or “No! Let me drown” could not plausibly be interpreted otherwise than as an act of refusing assistance. This is exactly where the drowning case falls short of systematicity. Recall that the condition requires that the recognition failure be connected to the illocutionary act the speaker tries to perform rather than to the linguistic form of the sentence uttered. A speaker is silenced if she cannot fully successfully perform a certain illocution no matter how she recasts what she said. One cannot escape from silencing just by changing the choice of words used: if the illocutionary act stays the same, one will generally be silenced all the same. While it suffices that Sally finishes the sentence for which her “No” was elliptical to overcome Peter’s failure, this route is precluded to the woman in the sexual refusal scenario: there is no straightforward linguistic addendum to her utterance capable of making its force transparent to the hearer. Both her “No” and Sally’s “No” fail (neither of them achieves uptake), but only the former is silenced in my sense.

To reiterate, I have argued that it is rather implausible that, upon hearing the words “No! I wanna die” or “No! Let me drown”, Peter would fail to grasp Sally’s intention to refuse his assistance. Against the backdrop of my analysis of the systematicity condition for silencing, Sally would not suffer silencing-as-uptake-failure. Note, however, that Peter might plausibly think that, although Sally is convinced that she wants to drown, this is not what her ‘deep self’ wants. He might think that deep down every living being wants to survive. If so, then Sally would be silenced in the sense of sincerity (i.e. true feeling) silencing.²⁸ This does not constitute a challenge for my account, since I take it that the notion of silencing captures very problematic cases (e.g. the silencing of women’s sexual refusals or expert speech acts) as well as less problematic or unproblematic cases (e.g. the silencing of Oliver’s act of telling; the silencing of the actor’s warning). Sally’s case would figure among the latter—that is, among those occurrences of silencing that do not track existing social disadvantages.

5 Conclusion

Illocutionary silencing is a multifaceted phenomenon. In this paper, I have first looked at the traditional conception of silencing as uptake failure and then I broadened the picture to account for other types of recognition failure. I have defined silencing as a linguistic interference brought about by a failure on the hearer’s part to recognize one or more constituents of the speaker’s illocutionary performance. Such an interpretation can be theoretically backed up by a social account of illocution. As I maintained throughout, the success of an illocutionary act is not only a function of the intentions of and the conventional means deployed by the speaker but is also partly dependent on how the act is recognized or taken up by the hearer. The hearer’s states of mind, that is, matter too. The ‘gap’ that separates the pronouncement of a certain utterance from the hearer’s recognition of its pragmatic components creates the possibility of silencing. Following suit with the literature, I have claimed that an instance of recognition failure constitutes silencing only if it is systematic. I distanced myself from the literature in taking this systematicity condition to require that the recognition failure be tied to the illocution the speaker tries to perform, rather than to the linguistic form of the sentence uttered. Finally, I submitted that, even though silencing often contributes to a broader unjust distribution of social power, structural injustice is not built into its very definition.