The pragmatic structure of refusal

This paper sets out to unpack the pragmatic structure of refusal—its illocutionary nature, success conditions, and normative effects. I argue that our ordinary concept of refusal captures a whole family of illocutions, comprising acts such as rejecting, declining, and the like, which share the property of being ‘negative second-turn illocutions’. Only proper refusals (i.e. negative replies to permission requests), I submit, require speaker authority. I construe the ‘refusal family’ as a subclass of the directives-commissives intersection. After defending my view against a number of potential objections, I highlight how a theoretically grounded analysis of refusal is not only of intrinsic value, but may also have significant moral and legal implications.


Introduction
Refusals are clear instances of illocutionary acts. This is shown by a number of factors. First, the verb 'refuse' can be aptly used performatively: one can refuse solely by saying that one is refusing. Secondly, and relatedly, utterances of the form 'I refuse to do ϕ' meet the so-called 'hereby criterion' (Austin, 1962, p. 57). An utterance of (1) I hereby refuse to authorize release of my health information would be perfectly appropriate. But an utterance of (2) * I hereby persuade you that the killer is the butler would be nonsensical: (2) cannot, of itself, persuade you about anything. In J.L. Austin's jargon, I cannot persuade you in saying things; I can persuade you by saying things. 1 Thirdly, refusals can succeed even if the causal effects they typically trigger do not obtain. I can succeed in refusing to authorize release of my health information, even if you ignore my refusal and disseminate my data; but I cannot succeed in persuading you that the killer is the butler if you do not come to believe it. This further confirms that the act of refusing is illocutionary in character.
Quite surprisingly, however, refusal is not listed in Austin's classification of illocutionary acts, and, in his work, nearly nothing can be found about refusing. This oversight has been partly remedied by social philosophers of language, who have recently emphasized that refusal, qua illocutionary act, can be easily performed under ideal circumstances and just as easily silenced under non-ideal ones. 2 Despite garnering philosophical attention, refusal remains largely underanalyzed. This paper sets out to unpack its pragmatic structure, by examining its illocutionary nature, success conditions, and normative effects. In Sect. 2, I argue that refusing has both a commissive and a directive facet. I also characterize refusals as 'negative second-turn illocutions', i.e. negative replies to some previous interrogative speech act performed by the hearer. In Sect. 3, I submit that our ordinary concept of refusal captures a whole family of illocutions, comprising acts such as rejecting and declining, which share some properties but also differ in important respects. In particular, while all family members are negative second-turn illocutions, only proper refusals (i.e. negative replies to permission requests) require speaker authority. This applies, mutatis mutandis, to the positive counterpart of refusal, namely, consent. The view on refusal I develop is subject to a number of potential objections, which I address in Sect. 4. I conclude by highlighting that, in addition to its importance for speech act theory, a systematic analysis of refusal may have significant moral and legal implications (Sect. 5).

A speech act analysis of refusal
We use the verb 'refuse' in several different ways. Some such uses do not capture any illocutionary act. Suppose I say that little Henry has refused to get intimidated by school bullies, intending to suggest that he did not let their threatening behavior scare him. In the (neo-)Austinian framework I endorse, illocutionary acts essentially reshape the normative landscape in such a way as to adjust people's rights, duties, and so on (see, esp., Sbisà, 1984Sbisà, , 2007. Henry's resilient attitude does not in itself affect the normative context: bullies have no right to threaten him regardless of how resilient he is. My use of 'refuse' does not pick out any illocution-and a fortiori any illocution of refusal. Since I set out to analyze the illocutionary act of refusal, I will leave cases of this sort aside. Other uses of 'refuse' capture illocutionary acts, but not necessarily illocutionary acts of refusal (as I conceive of it). Compare the following utterances.
(3) Larry has refused to take my money.
(4) A number of US citizens has refused to accept that Joe Biden won the 2020 election.
In (3), 'refuse' names an illocution by which one puts a halt to something; in (4), it names an illocution by which one rebuts a statement. I will argue that refusals have the normative (definitional) function of preventing a certain conditional obligation from turning into an unconditional one. When I offer Larry money, I acquire an obligation to give him that money provided that he accepts. In refusing, Larry prevents that conditional obligation from turning into an actual unconditional one. The election deniers' case is rather different, and I take it to involve a rebuttal, rather than a refusal. Rebuttals are assertions whose propositional content is contrary to (or incompatible with) some previously expressed proposition. 'Refusing to accept' that Joe Biden won the 2020 election amounts to asserting that Joe Biden did not win (or may not have won) the 2020 election, when the proposition that Joe Biden won the 2020 election has been previously asserted, suggested, or otherwise put forth. In what follows, I focus on refusals rather than rebuttals-on 'practical halts' as opposed to 'counter-assertions'. 3 This way of narrowing down the category of refusals is largely in line with Searle and Vanderveken (1985)'s analysis. They write: Just as one can accept offers, applications, and invitations, so each of these can be refused […]. A refusal is the illocutionary denegation of an acceptance [...]. One can only accept or refuse a speech act that allows for the option of acceptance or refusal (Searle & Vanderveken, 1985, p. 195).
The passage points at two core properties of refusal. Let us discuss them in turn.

Commissive directives
Searle and Vanderveken suggest that a refusal is the 'illocutionary denegation', or the negative counterpart, of an acceptance. 4 Like acceptances, refusals can be categorized as directives or commissives. If I accept that you do ϕ, I influence your behavior and hence perform a directive. 5 By contrast, if I accept (i.e. grant) your request that I do ϕ, I commit myself to doing it and hence perform a commissive. A general rule seemingly follows: interrogative commissives (such as offers) are replied to with directives; interrogative directives (such as requests) are replied to with commissives. The examples below may help illustrate the rule. 3 The Stalnakerian view is that an assertion is a proposal to update the common ground with what one asserts. (See, e.g., Stalnaker, 1999Stalnaker, , p. 10, 2002Stalnaker, , 2014.) I remain neutral on whether this is an appropriate analysis of assertion. But if it is, then the distinction between refusals and rebuttals may be less clear cut than I suggest here. Qua proposals, assertions would inherently call for acceptance or refusal on the hearer's part, and form a (sui generis) species of what I term 'open calls' (Sect. 2.3). 4 In Sect. 3, I claim that, strictly speaking, refusals are the negative counterparts of consents. At this stage, however, the distinction between acceptance and consent can be glossed over. 5 While Searlean directives are characterized as "attempts […] by the speaker to get the hearer to do something" (Searle, 1975b, p. 355), the directive class is traditionally taken to include also speech acts by which the speaker reshapes what the hearer is entitled to do. (See, e.g., Bach & Harnish, 1979, p. 47;Searle, 1975b;Searle & Vanderveken, 1985, p. 195.) It would be somewhat misleading to say that, in accepting that you buy me a drink, I 'attempt' to get you to buy me a drink. Similarly, in granting you permission to borrow my bike, I do not attempt to get you to borrow my bike. Yet, my speech acts can both be described as directives. After all, in granting you permission to do ϕ, I am influencing (or 'directing') your conduct by waiving an obligation not to do ϕ that you would otherwise have had.
However, things are not that simple, and the above is more of a guideline than a rule. For, interrogative speech acts typically have a hybrid nature, which is passed on to the yes-or-no response that follows. Consider the act of inviting. When I invite you to a Halloween dinner party at my house, I am trying to direct your conduct. 6 But more than that is involved. If I invite you to my party and, when the day comes, I refuse to let you in, you will normally have grounds to object. This is so because an invitation has also a commissive aspect: when I invite you to do ϕ, I am not only trying to make you do ϕ; I am also (perhaps, implicitly) committing myself to certain future behaviors. Invitations are commissive directives: hybrid speech acts that combine directive with commissive illocutionary force (Hancher, 1979). As such, they are followed by hybrid responses, which include a commissive component and a directive component designed to respond to invitations' directive and commissive facets, respectively. If you accept my invitation, you will be committing yourself to come to my party, but also implicitly directing my future conduct. Your affirmative response will, say, make me set a place for you at the dinner table; it will guide me to prepare some gluten-free dishes that I would not have prepared otherwise; it will make me let you in when you show up at my front door on Halloween; and so on. (The same would go if you refused my invitation: you would be committing not to come to my party, and I would consequently take a number of actions consistent with your refusal-e.g. I would set one less place at the dinner table.) Such remarks extend over a range of interrogatives and responses to interrogatives. Proposing, for example, has the same double nature. When I ask you for a dance, I am trying to direct your behavior, but I am also committing myself to a corresponding course of action. If you accept my proposal and I then refuse to dance with you, my (terribly rude) behavior will constitute a breach of commitment. Your affirmative response, on the upside, will commit you to dancing with me as well as typically influence my subsequent behavior.
As one can see, acceptances and refusals are midway between directives and commissives: sometimes, their directive nature is predominant (as in (5)); other times, their commissive nature takes over (as in (6)); more often, however, they are clearly two-faced (as in the Halloween party and the dance cases above). 7 6 Of course, invitations, like many other speech acts, can be performed insincerely: I can invite you to a Halloween party at my house out of politeness while hoping that you decline the invitation. Note, however, that, unless my speech act is openly insincere, in inviting you to come to the party I will still commit myself to preferring that you accept-and you will be justified in taking me to be trying to make you come to the party. I am grateful to an anonymous reviewer for suggesting that I discuss insincere invitations. See also fn. 10. 7 Austin's remarks on consent square with my analysis. In the twelfth lecture of How to Do Things with Words, the act of consenting-together with other permissives like authorizing and permitting-is classified both as an exercitive and a commissive. Austinian exercitives partly overlap with Searlean directives (Austin, 1962, 154ff;Searle, 1975b).

Second-turn illocutions
A second core property of refusals is that they cannot be performed in isolation. "One can only accept or refuse a speech act that allows for the option of acceptance or refusal", say Searle and Vanderveken (1985, p. 195). I can only refuse in reply to a speech act, and provided that it gives me the option to accept or refuse. Refusals are (what I call) second-turn illocutions, i.e. replies to some previous 8 speech act put forth by the hearer. 9 Their second-turn character complicates the analysis, which will have to preliminarily go through the first-turn illocutions refusals reply to.
What I label 'first-turn illocutions' are a subclass of what Mark Lance and Quill Kukla (writing as Rebecca Kukla) refer to as calls, i.e. second-person illocutionary acts that call upon the addressee to respond in certain ways (Lance & Kukla, 2013). Imperative (or closed) calls, such as orders or commands, oblige the addressee to do something. Interrogative (or open) calls, such as requests or invitations, give the hearer a reason to act in certain ways without obliging them to do so. In Lance and Kukla's way of putting it, open calls create petitionary reasons-they petition the addressee to do ϕ while leaving the decision to accept or refuse up to them (Lance & Kukla, 2013, p. 462). This is not to say that requests or invitations do not tip the balance in favor of acceptance. Quite the contrary: any interrogative call opens up a non-neutral space of choice. In inviting you to my party, for example, I try to make you accept the invitation. More generally, responses that fulfill the call's perlocutionary goal are preferred over, and thus more pressing than, responses that frustrate it. 10 Both accepting and declining are appropriate responses to an invitation, but the former is a more pressing option than the latter (Lance & Kukla, 2013, p. 461; see also Levinson, 1983, 307f).
So, interrogatives and imperatives differ in that only the latter impute obligations upon the addressee. To this, one might object that once an open call is accepted, an obligation typically springs into existence. If I request that you pick up our daughter today at school and you grant the request, then you will have committed to picking up our daughter today at school. However, the obligation on your part is not created by my request, but by your granting it. Petitionary reasons, I suggest, are akin to conditional obligations: obligations that get created only if the addressee answers affirmatively. Notice that the commissive-directive nature of both open calls and their replies makes 8 In non-standard cases, acceptances and refusals may precede the speech acts to which they reply. If I am confident that John is about to ask me to marry him, I can say "Yes" or "No" without even waiting for him to propose. Note that I perform a felicitous acceptance or refusal only if John later confirms, more or less tacitly, that he was about to propose indeed. If, to my enthusiastic "Yes", he replied, "Yes to what? I wasn't about to ask you anything", my acceptance would misfire. In more general terms, when we accept or refuse speech acts that we think our interlocutors are about to perform, our acts, if felicitous, are temporally prior to the 'first-turn' illocutions they reply to, which our interlocutors perform by confirming that they were about to perform those illocutions indeed. 9 In the parlance of conversation analysts, refusals are second parts to 'adjacency pairs', i.e. two-part exchanges in which the first part makes the second relevant and expected (Schegloff & Sacks 1973; see also Levinson 1983, p. 306). 10 There may be cases where a standardly preferred response is actually dispreferred by the speaker. Suppose that Julie throws a party for her 40th birthday and invites Madison, her annoying boss whom she despises, just because she feels she has to invite her. Even if in standard cases acceptance is the preferred response to an act of inviting, Julie much prefers that Madison declines her invitation. This divergence between standard and actual preferences is an exception explained by the caller's insincerity. it the case that the obligation in question partly (and sometimes primarily) falls on the caller. This is clear in the case of offers. If I offer to buy you a drink, your acceptance will primarily create an obligation for me to do what I offered to do.
Open calls are first-turn illocutions, whereas closed calls are not. Orders or commands call for compliance-for a behavioral response that fulfills the obligation the call has created. In contrast, requests or invitations call for an illocutionary response that typically takes the form of a 'yes' or 'no'. Refusals are second-turn illocutions performed in response to open calls. In what follows, I argue for the claim that the success conditions for refusals vary with the type of open call they respond to. In particular, I contend that the type of open call performed determines whether the refusal (or acceptance) act that follows is subject to an authority condition.

Open calls
If you invite me to dinner, I can accept or decline the invitation simply in virtue of being the addressee of your call. And if my housemate requests that I take out the trash, I am ipso facto entitled to grant or reject their request. This is also applicable to asymmetrical contexts. Suppose my boss asks me to complete a certain task. If they are genuinely requesting (rather than indirectly ordering), 11 then I will be entitled to refuse in spite of occupying a subordinate position. In light of this, one might maintain that, in order to rightfully accept or refuse an open call, a speaker must meet a mere addressee condition. 12 Addressee Condition: The speaker is the one called-or has been licensed to reply on their behalf by the one called.
We seem to have reached a conclusion: whenever one is the addressee of an open call, one can accept or refuse it regardless of the position one occupies. This, however, is not entirely true-for, to respond to certain requests, one must have some sort of authority. To see this, consider the following utterances.
(7) Would you take out the trash, please? (8) Can I borrow your laptop for a second?
The former is a simple request: in uttering (7), Matt (the requester) seeks to get Clare (the requestee) to do something. The latter is a request for permission: in uttering (8), Matt seeks to get permission from Clare to act in a certain way. A useful test to assess the nature of a given request (i.e. whether it is a simple or a permission request) is to examine which individual will be performing the activity represented in the proposition. Once a simple request is granted, the one who will perform the requested activity is the requestee. In contrast, once a permission request is granted, the one who will be given a right to perform the activity in question is the requester. We can also examine how both the requester and requestee relate to the activity at stake. As Monica Cowart points out, a request for permission is appropriate (and indeed required) when the activity at stake is in the requestee's jurisdiction (Cowart, 2004, p. 512). Since the laptop belongs to Clare, the activity in (8) is such that Matt cannot rightfully perform it without obtaining Clare's consent. Clare is entitled to give (or deny) him permission to use the laptop because she is its rightful owner, and the laptop thus falls under the scope of her authority. In contrast, the activity in (7) (i.e. taking out the trash) is not something over which the requestee has special control.
It follows that one is not entitled to grant or refuse a request for permission if the requested activity does not fall under the scope of one's authority. If the laptop were Anna's, Clare would have lacked the authority to give (or deny) Matt permission to borrow it. When it comes to accepting or refusing a permission request, a speaker must satisfy an authority condition.
Authority Condition: The speaker has authority over the activity represented in the proposition-or has been authorized to reply on their behalf by the one who has that authority.
So, refusals are negative replies to open calls. Generally, and with the exception of permission requests, in order to felicitously refuse an open call it is enough that the speaker is the one called. When it comes to requests for permission, however, the requestee must have authority over the activity at stake for their refusal to count as felicitous. Thus, the success conditions for refusal change as the first-turn illocutions it responds to change. More to the point, a refusal is an authoritative illocutionary act only if preceded by a permission request. There are several potential objections to this construal of the act of refusing. Before considering them, let me further detail my proposal.

Interim conclusion: the refusal family
Traditional speech act theory has it that two illocutionary acts are identical when they have the same success conditions. 13 If this is so, then no illocutionary force can survive a change in its success conditions, for changing those conditions means turning the original force into a different one. In this framework, it is hard to see how refusal may comply with different success conditions depending on the call it replies to, while remaining the very same act-namely, refusal.
My response to this is that the ordinary concept of refusal is applied to a multiplicity of illocutionary acts, which share the property of being negative second-turn illocutions and differ from one another in that they reply to different kinds of open calls. Furthermore, some of them require speaker authority, while others do not. We can and do say, "I refused the offer", "I refused the invitation", "I refused to grant permission"; but the acts we accomplish in refusing an offer, an invitation or a permission request are not the same. Rather, they are different acts belonging to the same illocutionary family-the refusal family.
Even though there is no one-to-one correspondence between performative verbs and illocutionary acts, 14 we can give it a try and check for English verbs marking acts in the refusal family. Some verbs come easily to mind; for example, 'reject' and 'decline'. That rejections and declinations are part of the family is proven by the fact that they share the fundamental family property-i.e. 'being a negative second-turn illocution'. This means that they constitutively block the creation of certain obligations: any refusal act makes the obligation conditionally produced by the open call that preceded it fade away. At the same time, they have distinctive preparatory conditions specifying which kind of open call they respond to. Rejections typically respond to proposals and simple requests; speech acts of declining, by contrast, follow invitations.
The take-home message here is that we do many different things with the word 'no'. In saying "No", we can reject a proposal, decline an invitation, turn down an offer, and so on and so forth. We perform a proper refusal, I submit, when we deny a request for permission. Given the nature of permission requests, proper refusals, unlike other family members, require speaker authority. 15 What I have said for the word 'no' also applies to the word 'yes' and to the acts it typically realizes. Just as one can conceive of refusal as an illocutionary family, so too one can conceive of consent as a family comprising the positive counterparts of the acts we have been concerned with. All members of the consent family share the property of being affirmative second-turn illocutions. This normatively amounts to providing the condition under which the obligation conditionally produced by an open call is created. As with the refusal family, the members of the consent family differ from one another in virtue of the type of open call they reply to. In saying "Yes", one can agree to a proposal, accept an invitation, take an offer, etc. One performs a proper consent only if one thereby replies to a request for permission. 16 Given the nature of permission requests, proper consents, unlike other family members, require speaker authority.
In talking about illocutionary 'families', I am following Mitchell Green's terminology. Green (2013Green ( , 2017 argues that assertions comprise a family of speech acts whose performance commits the speaker to the truth of a proposition. 17 Examples include proper assertions, conjectures, guesses, and so on. Even though Green does not put it this way, the assertion family forms a subclass of assertives. The family leaves out those assertive acts-such as supposing a content for the sake of argument-that do not commit the speaker to truth. I argued above that consents and refusals are midway 14 Consider, e.g., the verbs 'assert' and 'state', which name the same assertive force while being nonsynonymous. See Searle (1975b, p. 345) and Searle and Vanderveken (1985, p. 179). 15 I agree with Mary Kate McGowan that «it is intuitively clear (and perhaps even obvious) that refusals concern permission» ( McGowan, 2009, p. 489). This is why I call negative replies to permission requests 'proper refusals'. 16 That the essential nature of consent is to grant permission is largely agreed upon. See, e.g., Cowart (2004, p. 514): «The act of giving your consent revolves around willingly giving permission to someone to do something that they do not have a right to do without asking for your permission»; Kleinig (2010, p. 20): «Consent as a communicative act must be intended to convey to B a permission or entitlement»; and Dougherty (2015, p. 226): «Consent is "morally valid" when, all else equal, it succeeds in generating a moral permission». 17 See also Sbisà (2019). between directives and commissives (Sect. 2.1). Although some of them may primarily be classified as directives (e.g. taking/turning down an offer) and some others as commissives (e.g. granting/rejecting a simple request), all of them typically pursue both a directive and a commissive point-i.e. they aim at directing the audience's conduct, but also constitutively commit the speaker to a future course of action. The refusal family and the consent family form subclasses of the directives-commissives intersection.
Before moving on, let me stress that little hangs on the terminological dispute over how we ordinarily use the verbs at issue. We may well say or hear people saying that a given proposal has been consented to or that a certain invitation has been rejected. My concern is not primarily with how people actually use these verbs; rather, the point for my purposes is that such verbs (and in principle other verbs as well) can be deployed to capture discrete illocutionary acts, whose pragmatic structures have family resemblances but are not identical.
In the next section, I consider some potential objections to my analysis of refusal. For the sake of simplicity, I will often talk about refusals in general. I will however adopt a more fine-grained terminology-and refer to rejections, declinations, proper refusals, and so on-when relevant.

Potential objections
The view on refusal I have advanced can be stated as a conjunction of three related claims: (i) refusals are second-turn illocutions; (ii) refusals are performed in response to open calls; (iii) depending on the nature of the call they respond to, refusals comply with different success conditions. Specifically, a refusal is an authoritative act only when preceded by a request for permission (that is, only when it is a proper refusal). This proposal may seem unsound to some. In this section, I consider a number of objections to each claim, and hope to show that, on reflection, the analysis is actually sound.

Refusals in isolation
For one thing, one might object to the claim that refusals are second-turn illocutions and contend that they may be accomplished in the absence of any previous call by the hearer. Consider the scenario that follows.
The Exhausted Intern: James works as an intern in a renowned law firm. Every workday starts the same way, with his boss Rhonda asking him to get coffee for her and James fetching coffee, as she asked. Until one morning, when as soon as he sees Rhonda and without even waiting for her to issue the request, James says a loud "No". 18 James' "No", one might argue, is a refusal performed in isolation, for Rhonda had not issued any request that day (at least, not yet). I claim that, despite appearances to the contrary, James' "No" is a second-turn illocution, for it replies to a habitual request on the part of his boss. Upon repetition, the request/granting pair becomes a default score component of their morning conversations. James' annoyed refusal does not occur in isolation, but in place of the second part of the pair, which by default would otherwise have been a granting of the request.
I borrow the notion of score from Lewis (1979). Roughly, conversational score is a register tracking all those elements that together determine what counts as appropriate in the ongoing conversation-sets of presuppositions, standards of accuracy, facts about salience, and so on. The score of routine conversations, I submit, also tracks shared information about the speech acts that participants habitually perform. Once that information enters the score, one may secure uptake for those speech acts and achieve their perlocutionary objects without saying a word. If I have breakfast at the same café every morning, and every morning I ask for a cappuccino and an apricot croissant, it is very likely that one day the barman will start to serve me a cappuccino and an apricot croissant without me even asking. Similarly, if Rhonda asks James to get coffee for her every morning, and every morning James grants the request and gets her coffee, at a certain point she will not have to ask anymore for James to show up with a steaming cup of coffee. The request/granting pair will have become a default score component of their morning conversations. In order to change it, one of them must perform a speech act that is incompatible with the default score. It might be Rhonda saying that she has banned coffee from her diet, or James refusing to get her coffee. By triggering the rule of accommodation, 19 moves like these change the score in such a way as to wipe out the default request/granting pair. The upshot is that James' refusal is performed in response to some previous open call-an open call that lies in the score as a default component rather than a newly set component.
Elinor Mason (forth.) has developed a different sort of case meant to prove my first claim (i.e. refusals are second-turn illocutions) wrong. The case concerns the Scottish habit of breaking into singing Auld Lang Syne in crowds.

Auld Lang Syne: It's New Year's Eve. Jason is spending the night in Princes
Street along with hundreds of other people. At the stroke of midnight, a rousing chorus of Auld Lang Syne breaks out, and everyone joins hands with the person next to them to form a circle. Except for Jason, who retracts hands while shaking his head. 20 It is very likely that the person next to Jason will not utter any words or make any gestures to propose that they hold hands. They will simply assume that Jason goes along with it and take his hand. Yet, Jason «can refuse: a refusal is an appropriate move» (Mason, forth., p. 7)-which would contradict my view that refusals can only be performed in response to a call by the hearer.
The way I see it, Jason responds to a tacit or implicit call. In Scotland, singing Auld Lang Syne on New Year's Eve has become a ritual to bid farewell to the old year. The ritual includes crossing your arms and holding hands with those who stand next to you. Once you decide to participate in the ritual chant, once you choose to play the game, you therewith accept its rules. You can of course withdraw acceptance later, even though this, in a sense, makes you no longer a player. When Jason retracts his hands, he withdraws the acceptance he had tacitly given the moment he started to sing along with the others. He turns that acceptance into a refusal. The refusal act he performs does not occur in isolation, but in response to an implicit proposal built into the rules of the game. To say this with Lewis, when people start singing Auld Lang Syne, the score already includes as many proposal/acceptance pairs as there are singers. To overturn one or more of those pairs, some participants must reject to hold hands. A partner-switching dance is a comparable case. Once you start dancing, you tacitly accept the switch. If you do not withdraw acceptance (and quit the game), then when the music stops you are expected to switch partners. The closest person next to you will not ask if you want to dance with them, because the proposal is already built into the game's rules-as is your acceptance, unless you retract it. 21

Refusals in response to closed calls
Secondly, one might agree that refusals are second-turn illocutions, but claim that they can also occur in response to closed calls. After all, just as one can refuse to do what one is requested to do, one can also refuse to do what one is ordered or even commanded to do. Before considering this objection, let me clarify how I conceive of the difference between orders and commands.
Both commands and orders are authoritative speech acts: their performance requires that the speaker have, and be exercising, authority. However, while commands require official authority (e.g. one must have a higher army rank to command a Private to do this and that), orders need informal speaker authority. A gang leader does not occupy any official position of authority; and yet, they can order their men to do such and such. The gang leader's informal authority enables them to successfully give orders. Expanding upon Ishani Maitra's idea of 'licensed' authority, Rae Langton has recently argued that a speaker's informal authority may spring into being thanks to their audience's compliance. 22 Orders, in her view, abide by a disjunctive authority condition-an order is successfully performed only if (i) the speaker has some sort of authority over the hearer before the time of utterance; or (ii) the hearer grants the speaker authority by letting their act pass. Commands, by contrast, abide by a simple authority condition requiring the speaker to have pre-established official authority.
With this in mind, let us consider two apparent counterexamples to my second claim (i.e. refusals can be performed only in response to open calls). The first involves a command; the second involves an attempted order. If what I have said in Sect. 2.2 about closed calls is correct, then Private Freeman should be given no possibility to refuse. Yet, she can-and indeed she does, in this fictional example-answer, "No, I won't do that". The Private refuses to obey the command; but a refusal to obey is not an illocutionary act of refusal. Recall that a successful closed call imputes an obligation upon the addressee. Insofar as the addressee acknowledges the call's successfulness, they have already acknowledged their obligation to obey. No acceptance is required for the obligation to be created-which implies that no refusal can come in the way of that creation. In the case of requests (and other open calls), by contrast, the obligation is conditional: it gets created only if the hearer grants the request. If Clare rejects Matt's request that she take out the trash, then she will not acquire any obligation to do so. The act she performs when she says, "Sorry, I can't", falls within the refusal family exactly because it blocks the creation of an obligation on her part. This explains why the insubordinate Private does not refuse. 24 Even if she says "No", the obligation gets created. She ought to go and pick up wood. If she does not, she will transgress an obligation. So, what kind of illocution does the Private perform with her "No"? My answer is that she performs an announcement of disobedience. In uttering "No", the Private is letting the General know that she does not intend to go and pick up wood. 25 Note that I am in no way denying that people can 'refuse to obey': of course they can. Transgression is an option. When we are successfully commanded (or ordered) to do something, we acquire an obligation to do it, which we can transgress. Sometimes, we transgress obligations without announcing that we will. (Imagine that General Hooks commands Private Freeman to go and pick up wood and that Freeman simply does not do that.) Other times, however, we announce that we will transgress a certain obligation. This is what Freeman does in the Insubordinate Private case when she replies, "No, I won't do that". What we ordinarily call 'refusals to obey' are, in my view, announcements of disobedience (or announcements of transgression). They are not illocutionary acts of refusals, for they do not have the normative function to prevent a conditional obligation from turning into an unconditional one.
What about orders? We saw that they do not require any formal speaker authority. Does this open up space for a refusal move on the addressee's part? To answer this question, it may be helpful to consider a revised version of our previous example.
The Unaccommodating Friend: Giulia and Lucy go camping in the Peak District. When they arrive, Giulia starts to give Lucy orders. "Go and pick up wood", says Giulia. Lucy replies, "No, I won't do that".
The difference between this and the previous scenario is that Giulia, unlike General Hooks, has no pre-established authority, and her order can only succeed if the addressee 24 The fact that we cannot refuse a closed call is mirrored in the way we speak English. As Searle and Vanderveken (1985, p. 195) point out, one cannot say that one refuses an order (or a command) but rather that one refuses to obey it. But with open calls, such as offers or invitations, one can literally say, "I refused the offer" or "I refused the invitation"-though we saw that offers are strictly speaking turned down and invitations declined. 25 Depending on how we construe the scenario, the Private might also be performing an act of challenging the General's authority (and perhaps, the whole hierarchal structure of military rank); but anyhow, she would not be performing any act of refusing. grants her authority on the fly by letting her act pass. When Lucy says "No", she is not refusing; nor is she primarily letting Giulia know that she has no intention to go and pick up wood (though this is something she may secondarily do). Lucy is primarily objecting to Giulia's bossy behavior. Lucy's resistance prevents Giulia from acquiring authority, thus making her order misfire. Another way to put it is to say that, thanks to Lucy's 'unaccommodating' reply, Giulia's attempt to order remains just that: a mere attempt, which fails to count as a felicitous order. 26

Refusals are authoritative illocutions
Thirdly, one might claim that refusals need speaker authority regardless of the call they respond to. In other words, rejections, declinations, and the like, would be authoritative illocutions just as proper refusals.
One way to cash out this objection is to argue that the addressee condition I put forth in Sect. 2.3 is an authority condition in disguise. Recall that the addressee condition says that one can felicitously reply to an open call only if one is the one called or has been licensed to reply on their behalf by the one called. But this can be read as telling us that any open call grants its addressee the authority to either accept or refuse it. Thus, so the argument goes, all refusals require speaker authority, and whether one satisfies the addressee condition is an authority issue after all. 27 It is quite clear, I think, that, if I offer you a coffee and someone else steps in and turns down the offer for you, their refusal will fall flat. My offer entitles you to reply. If this entitlement can be cast in terms of authority at all, it can be cast in terms of conversational authority: my offer grants you the authority to make a certain move in the ongoing conversation-precisely, to give a reply. When it comes to offers, 'conversational authority' is all that is needed for the addressee to felicitously accept or refuse. When it comes to requests for permission, however, 'conversational authority' is not enough. Clare can grant (or deny) Matt permission to borrow her laptop because, as its owner, she has authority over it, i.e. a right to dispose of it and loan it to others. And I can grant (or deny) my son permission to have a sleepover because, as his parent, I have authority over him, i.e. a right to make certain decisions concerning his upbringing and well-being. As one can see, authority, in these cases, goes well beyond an entitlement to make moves in the ongoing conversation. I conclude that, even if my addressee condition could be recast in terms of authority, the authority required to reply to offers (and the like) and the authority required to reply to permission requests would be substantially different, and my overall analysis would stand.
One might, however, cash out the objection in a different way. Consider a case developed by Mason (forth.).
The Man under House Arrest: Amanda invites Brian to a party at her house. Brian refuses. Unbeknownst to Amanda, Brian is under house arrest, so he did not have the authority to accept-which makes his refusal non-ideal or defective. 28 Thinking about exceptional cases like this reveals, Mason claims, that 'extraconversational' authority is required to decline invitations, even though at first glance it might not seem so. Mason puts forth a three-part argument: (i) Brian cannot fully successfully accept Amanda's invitation; (ii) this is so because Brian lacks the requisite authority; (iii) since Brian does not have the option of accepting, he cannot fully successfully decline the invitation either. In what follows, I will argue against each claim in turn.
As already said, accepting an invitation has a commissive facet: it commits the speaker to participating in the event they have been invited to. All commissive forces have the common preparatory condition that the speaker can do what they commit themself to doing. A speaker who accepts to do ϕ presupposes that they can do ϕ. Suppose Brian accepts Amanda's invitation. In doing so, he presupposes that he can be at her place the day of the party. Following Mason's line of thought, one might conclude that, since Brian cannot be there (recall that he is under house arrest), his acceptance is non-ideal or defective. Such a conclusion, however, would not be supported by classical speech act theory, which interprets the general preparatory condition for commissives stated above in terms of physical possibility, and not of legal or otherwise deontic possibility. 29 One underlying reason is quite obvious-namely, people can choose to break a deontic rule they are subject to, but they cannot equally choose to break a physical law. In the real world, Scotty cannot promise Captain Kirk to beam him up, because it is physically impossible to teleport a material object. Given the laws governing the motion of matter, and the rules defining a successful promise, Scotty's speech act would inevitably go awry. But with deontic laws, things are rather different. In being sentenced to house arrest, Brian acquired the obligation not to leave his house until he has served his sentence. Such an obligation imputes deontic constraints on Brian's behavior that Brian can choose to violate. One may claim that any speech act performed in breach of an obligation is at best defective, but I think that this is rather counterintuitive. Imagine someone is offered an illegal drug. It seems to me that one could either take or turn down the offer, and do so fully successfully, no matter what the law prescribes. 30 Of course, one might insist that, albeit counterintuitive, it is indeed the case that the speaker's deontic status-their package of rights and duties-impacts on the range of speech acts that they can (fully successfully) perform. If so, Mason's claim (i) may turn out to be true, but not for the reasons she provides. Brian's impossibility to fully successfully accept Amanda's invitation would have nothing to do with the scope of his practical authority. To make this point vivid, consider an alternative scenario.
The Plainclothes Officer on Duty: Mark offers to buy Robin an alcoholic drink. Robin refuses. Unbeknownst to Mark, Robin is a plainclothes officer on duty, so she was not allowed to accept-which makes her refusal non-ideal or defective.
In the perspective we are dealing with, had Robin said "Yes" and taken the drink, she would have performed a defective illocution; but this time, it is rather clear that her act's defectiveness would not derive from a lack of authority. Mason's argument is intuitively sound because the example she chooses involves home detention, and home detention makes the scope of one's authority smaller. But this does not mean, not per se, that a detainee's impossibility to accept a certain invitation is an authority issue. Mason's claim (i.e. authority is involved in the acceptance/refusal of any open call) loses its intuitive soundness as soon as we slightly modify the example. When we consider a police officer's impossibility to accept an alcoholic drink, we are not that convinced that this has to do with a lack of authority. For, when one joins the police, the scope of one's authority gets overall bigger. Conceding that Robin's "Yes" would constitute a defective illocution, this would be so not because she lacks the authority to accept, but because her speech act would violate a preparatory condition requiring that it is deontically possible for the speaker to do what they accept to do. So, even taking on board Mason's claim (i), her claim (ii) would not appear to be compelling.
Finally, I have claimed that refusals are the negative counterparts of acceptances, with which they share crucial properties (first and foremost, they both are second-turn illocutions). This does not mean, however, that they have the same success conditions. To accept an invitation is to commit oneself to doing ϕ (e.g. to showing up at a certain party). In contrast, to refuse an invitation is to commit oneself to refraining from doing ϕ (e.g. to refraining from showing up at the party). While accepting an invitation requires one to be able to do ϕ, refusing an invitation requires one to be able to refrain from doing ϕ. When Brian replies, "Sorry, I can't", to Amanda's utterance of "Would you like to come to a party at my place next Friday?", he performs a fully successful act, regardless of whether an acceptance on his part would have been successful. For legal reasons, Brian cannot participate in the party, and he is saying just that-I can't.
To recap, I have suggested that a speaker can fully successfully accept an invitation even when doing so is at odds with some deontic rules they are subject to. One might insist that there is a sense in which accepting an invitation to a party when one is under house arrest, or accepting a drink whilst on duty, is to perform a defective illocution. Even if so, I have argued that the defectiveness of those acceptance acts would not depend on a lack of authority on the speaker's part, as Mason contends, but on the set of constraints generated by the speaker's pre-existing duties. I have finally maintained that Mason's argument unduly conflates the success conditions for accepting and refusing. Even though such speech act types are closely related (they stand in a relationship of illocutionary denegation to one another), the former requires that one can-physically or deontically-do what one accepts to do, whereas the latter requires that one can refrain from doing what one refuses to do. And this can make it the case that a refusal is fully successful when an acceptance would not be.

Refusals are not authoritative illocutions
Lastly, one might contend that refusals do not require speaker authority, not even if preceded by requests for permission. Recall that my third claim was that refusals are authoritative illocutions only when preceded by permission requests. This claim has two main rivals-it can be challenged by arguing either that authority is always needed to refuse or that it is never needed to refuse. I have dealt with the first rival claim in the previous section. Here, I take into account the second. Let me introduce it with an example.
The Wrong Person: Matt needs to send an email. Seeing a laptop on Clare's desk, Matt asks her permission to borrow it. Clare, however, is the wrong person to ask, for the laptop is Anna's. This makes Matt's request void, and consequently, any reply on Clare's part flawed.
Permission requests are successful, one may argue, only if addressed to the right person. Since Matt addresses the wrong person, his act misfires. Clare cannot successfully answer yes or no because Matt's request is void-not because she lacks authority. 31 I take issue with the claim that Matt's request misfires. Requests for permission, I suggest, are governed by a preparatory condition requiring (i) that the hearer has authority over the activity in question, or (ii) that the speaker believes that the hearer has authority over the activity in question. If both (i) and (ii) are fulfilled (and provided that further conditions hold), then the request is fully successful. If both (i) and (ii) are not met, then it is a misfire. But if only one disjunct obtains, then the act is less than fully successful (or successful though defective). Matt's call does not satisfy (i) (for Clare has no authority over the laptop in question), but it satisfies (ii) (Matt believes that Clare has that authority). Matt's call is therefore defective, not a complete misfire.
At this point, an opponent might concede that Matt's request partially succeeds, but insist that it is misplaced, and that this is what we should focus on when we assess the success of a yes or no on Clare's part. Clare's authority does not enter the picture. While I agree that Matt's request is misplaced (he should have addressed Anna rather than Clare), I do not see how Clare's authority can be taken out of the picture. Matt's request is misplaced exactly because Clare lacks authority over the laptop. To see this, it may be helpful to compare the laptop case to a similar case involving an invitation rather than a permission request.
The Clumsy Professor: Professor X, Professor Y, and Professor Z are at a conference coffee break. Professor X wants to invite Professor Y-who he knows by reputation-to deliver a seminar at his University the following semester, and has prepared an official invitation letter. However, Professor X is a little clumsy and, not knowing how Professor Y looks like, ends up giving the invitation envelope to Professor Z. 32 Professor X's invitation is clearly misplaced; he intended to address it to a different person than the actual addressee. Now, suppose that the following day, Professor Z opens the envelope and sees that the heading reads 'Dear Professor Y, …'. Professor Z immediately realizes the mistake, and that she cannot accept or refuse the invitation. Notice that no authority issue is involved here. The fact that Professor Z cannot accept or refuse the invitation derives from the fact that she is not the intended addressee of the call, even though she happened to be the actual one. One might claim that the same goes with misplaced requests for permission, but I do not think so. Professor X wanted to invite Professor Y, but he could have wanted to invite Professor Z. For, which person one invites to a certain event is largely a matter of choice. 33 On the contrary, in the laptop case, Matt could not have asked anyone but Anna. For, which person one asks permission to do something is not a matter of choice. It is a matter of authority-of the authority that certain persons, and not others, have over a certain activity.

Interesting practical implications
In addition to its relevance for speech act theory, the analysis I carried out may have consequences for issues of moral and legal concern. I will just give one example.
Some theorists have claimed that sexual refusals and sexual consents are authoritative illocutions. In saying "Yes" or "No" to sex, one exercises authority over one's own body. 34 While the claim may seem unproblematic, my analysis suggests that it is true only if sexual advances are requests for permission. And this is problematic. Common requests for permission involve either possessions or guardianship issues. Clare can grant Matt permission to borrow her laptop because it is her laptop (and hence she has disposition rights over it), and I can grant my son permission to have a sleepover in virtue of the parental authority that I have over him. Both permission request types typically assign a role of passivity to the requestee. Once Clare consents to letting Matt borrow her laptop, Matt gets to borrow her laptop, but Clare, for her part, does not get to do anything. Once I consent to my son having a sleepover, it is him-not me-that gets to sleep away from home. And if a student asks me if they can have an assignment extension, it is the student that gets a few extra days to turn in the assignment in case I grant their request.
This gives some support to the claim that a person's (proper) consent extends the latitude of another person's agency; but for their own part, the consenter makes themself "more or less a patient" (Gardner, 2018, p. 58). Applying the category of permission request to sexual advances turns out to be highly controversial: it is to offer in backlight a view of sex whereby one party's sexual agency finds expression, while the other party, to an extent, makes themself a patient. 35 Giving a full argument against a permission request model of sexual advances goes well beyond my current purposes. 36 I just want to point out that, if one abandons the model, then one must also abandon the view that sexual refusals and sexual consents are authoritative illocutions. 33 I acknowledge that this is oversimplistic and that complex considerations go into deciding whom to invite, say, to one's wedding or to certain work meetings. My point here is not that invitation lists are always at our full discretion. I am more modestly claiming that we typically have some discretionary wiggle room over whom we invite to our events-room we do not have when it comes to requests for permission, which need to be addressed to whomever has authority over the activity at stake. I thank an anonymous reviewer for pressing me to make my point clearer. 34 See, e.g., McGowan (2009) andSbisà (2009). 35 Critics of the language of consent in sexual ethics have contended that consent implies passive acquiescence on the part of the consenter. (See, esp., Anderson, 2005;Gardner, 2018;Kukla, 2018;MacKinnon, 2016.) I suggest that this is so because of the nature of permission requests, to which acts of consent reply. I further develop this point in Caponetto (2017). 36 Kukla (2018) gives a much more detailed argument, in terms that are broadly compatible with the argument I sketch here. See also Caponetto (2021).
More than that, if, generally speaking, sexual advances are not permission requests, when we talk about sex, 'consent' and 'refusal' are not even the right categories.
In this paper, I have put forth and defended a speech-act analysis of refusal, according to which refusals are second-turn illocutions performed in response to open calls and requiring speaker authority only when preceded by requests for permission. I have maintained that our ordinary concept of refusal captures a family of illocutions, which lie at the directives-commissives intersection and whose normative function is to prevent the creation of an obligation. Open calls produce conditional obligations-obligations that become binding only if the addressee accepts. Refusing is an illocutionary tool to avoid incurring obligations we are not willing to shoulder. 37

Conflict of interest
The author has no relevant financial or non-financial interests to disclose.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/ by/4.0/.