Permissive consent: a robust reason-changing account

Abstract

There is an ongoing debate about the “ontology” of consent. Some argue that it is a mental act, some that it is a “hybrid” of a mental act plus behaviour that signifies that act; others argue that consent is a performative, akin to promising or commanding. Here it is argued that all these views are mistaken—though some more so than others. We begin with the question whether a normatively efficacious act of consent can be completed in the mind alone. Standard objections to this “mentalist” account of consent can be rebutted. Here we identify a much deeper problem for mentalism. Normatively transformative acts of consent change others’ reasons for acting in a distinctive—“robust”—way. Robust reason-changing involves acts aimed at fulfilling a distinctive kind of reflexive and recognition-directed intention. Such acts cannot be coherently performed in the mind alone. Consent is not a mental act, but nor is it the signification of such an act. Acts of consent cannot be “completed” in the mind, and it is a mistake to view consent behaviour as making known a completed act of consent. The robust reason-changing account of consent developed here shares something with the performative theory, but is not saddled with a label whose home is philosophy of language. Certain kinds of performative utterance may change reasons robustly, but not all robust reason-changing involves or requires acts of speech, and consent can be effected by a wide range of behavioural acts.

Adult individuals can render permissible, by acts of consent, actions that would otherwise be impermissible. There is a debate about what kind of thing consent must be, if it is to bring about this kind of normative change. One view—mentalism—is that performing the appropriate kind of mental act is not just necessary but also sufficient to bring about the normative change distinctive of permissive consent (Hurd 1996; Alexander 1996; Husak 2006; Alexander 2014; Dsouza 2013).¹ Here our focus will be upon the sufficiency claim: can a merely mental act be sufficient for the kind or normative change that consent brings about?

Mentalism may seem obviously false and not to need refutation: how can performing an act in the mind directly change other people’s reasons for acting? However, there are arguments in favour of it, and a rejection of mentalism is not as straightforward as it may seem. The arguments standardly offered against mentalism fail to acknowledge the distinctive assumptions which motivate and sustain mentalism. But mentalism is false, and the aim here is to show why. The mentalist has to accept that consent is typically made via intentional, overt, behavioural acts, and thus has to have something to say about the relationship between the normatively efficacious mental acts and overt consent behaviour. Mentalism views such behavioural acts as playing a primarily epistemic role: that of making known, to relevant others, that the normatively transformative mental act has been made. However, by reflecting on the distinctive way that the exercise of normative powers changes other parties’ reasons for acting—robust reason-changing—we see that this assumption about the role of consent behaviour is misguided.² Overt behavioural acts of consent are not aimed at making known an act of consent that has already been “completed” in the mind. Such acts are aimed at fulfilling a distinctive kind of reflexive, recognition-directed, intention. The robust reason-changing nature of consent implies that acts of consent cannot be effected in the mind alone. The argument offered here also applies to the view that consent is a “hybrid” of some kind of mental act plus some further act which signifies, expresses or communicates that act.

1 Mental necessity

Our focus is upon permissive consent, rather than, say, consent as agreement, or consent to be bound by terms of a contract. Our focus will also be restricted to individual consent, rather than permission given by institutions, collectives or groups. Permissive consent of this kind is operative against a backdrop of conditional norms of the form: S may not do X, unless R consents. R’s act of consent renders the doing of X permissible for S. Permissive consent is operative with regard to different kinds of norms: moral norms; legal norms; social norms, norms of politeness, and norms established by contract, or agreement, for example. Here our focus will primarily be on consent as morally transformative.

There are a number of reasons to hold that certain kinds of mental element, including mental acts, are necessary for normatively transformative consent.

1.1 Discretionary powers and the normative significance of choice

The power to permit by consent is a discretionary one. It is up to the agent herself whether to exercise her power, or not. She may do so on a whim, or for cruel selfish reasons, or against her own best interests. To whom the individual gives consent, and for what is also up to the individual herself. Not only is consent essentially mind-involving because it is discretionary—and thus requires a decision on the part of the consentor—it is arguable that the reason why individuals have a discretionary power—in whatever normative context consent is operative—is that choice has a normative significance (in that context). For example, in the moral realm it is arguable that we have a discretionary power to permit by consent because we have a fundamental interest in being able to decide whether, when, and for whom, to create an exception to certain of our moral rights (Scanlon 1988; Owens 2011). Thus, in many moral consent domains—such as medical treatment and research, and sexual contact—the power to permit by consent is explicitly justified by an appeal to respect for autonomy (Childress 2012).³ To perform a medical intervention, or engage in medical research, without consent is to fail to respect the patient, or research subject, as an autonomous person, capable of making her own decisions about what to do, or what is to be done to her.⁴ Similarly, theories of what makes non-consensual sex morally wrong make appeal to our rights as individuals to autonomously decide who is permitted to make certain kinds of contact with us (McGregor 2005; Archard 2007).

1.2 Content and normative scope

Consent is directed at certain kinds of action, and certain agents. It is (typically) directed at acts that have not yet been performed.⁵ The acts to which consent is directed need not actually be performed. But there is nothing odd about this relationship to not-yet-performed actions once we acknowledge that consent has an essential mental component: consent can be directed at acts that are never performed just as intentions to act do not entail the performance of the act.

Not only do acts of consent have an intentional content—of the form S may do X (where S can be an agent, or set of agents, and X a type of action)—acts of consent have a normative scope. Consider social norms of politeness. S asks R if she may take the vacant seat on the train beside R. If R consents (leaving aside for now exactly what it is that R must do), her consent pertains to S, and to the action of taking the seat. It does not cover T’s taking the seat, or S’s borrowing R’s newspaper. The same point is true in other normative contexts. If R consents to S taking her vintage motorbike, if T then does so (without consent being directed at T), then T wrongs R. Similarly, if S takes R’s prize collection of cutlery, she wrongs R. In each case the act of consent only renders permissible some actions for some agents. The content of an act of consent is related to, but not the same thing as, its normative scope. The content of an act of consent is what it would permit, provided that certain conditions are met and it is not undermined by “defeating” conditions. The normative force of an act of consent—whose intentional consent is that S may do X—can be defeated, such that even though the act is made, the result is not that S may do X.

1.3 The contrast between transformative and inert acts of consent

Focusing on the defeating conditions for consent gives us another line of support for the claim that certain kinds of mental act are necessary for normatively efficacious consent. Consider the contrast between the normatively transformative PERMISSION and the two normatively inert examples that follow it:

• PERMISSION: S asks R if S may kiss R’s hand. R nods and offers her hand to S.

• THREAT: S tells R that if she does not let him kiss her hand he will kill her. R takes S’s threat to be credible. R nods and offers her hand to S.

• IMPERSONATION: S impersonates R’s lover T. Pretending to be T, S asks to kiss R’s hand. R thinks that it is T is asking. R nods and offers her hand to S (believing him to be T).

The latter two examples are ones where R performs the same type of behaviour—offering her hand to S—as she does in the permissive case, but, intuitively, in the latter two situations the act performed does not bring about normative change. Such an act is normatively inert, at least with regard to the kinds of normative change distinctive of consent.

There are different ways that we might explain this contrast, but one plausible way is that in the inert cases R lacks the right mental state, or fails to perform the right mental act: that is, the kind of mental act, with the content necessary for normative change.

What kind of mental state or act? This is a question that we will answer in more detail below. Let us first consider what mentalists assume about the kind of mental state that is necessary—and, in their view, sufficient—for normatively transformative consent. Heidi Hurd, argues that normatively efficacious consent involves R intending that the act in question be performed.⁶ This may seem to have some plausibility. In THREAT and IMPERSONATION R does not intend that the action that is in fact performed, be performed. But there are two problems with this line of thought. First, one can permit an action without intending that it be performed. Consent can be given reluctantly, as the “least bad” option from a poor set of options. Second, the very idea of intending someone else’s actions is puzzling: R can hope that S kisses her hand, or wish that he would, but she cannot directly intend that he does so (Alexander 2014). Larry Alexander offers a more plausible candidate:

The mental state that I believe constitutes consent is that of waiving one’s right to object, or, if that sounds too much like a non-mental action, that of mentally accepting without objection another’s crossing one’s moral or legal boundary (the boundary that defines one’s rights. (Alexander 2014: 7)

Although Alexander talks of the “mental state” of consent, he is clear that consent is an act, not some state, condition or disposition. Consent is not some kind of state that might befall the consentor.⁷ In the normatively inert examples above, R does not mentally waive her rights against S doing the action in question, even though she behaves as if she does.

2 Mental sufficiency

Suppose we accept that certain kinds of mental act are necessary for normatively transformative consent, can a mental act be sufficient for normative change? Hurd (1996) asks us to consider an example of a severely disabled person incapable of communicating by any means. Hurd argues that such a person can still give normatively transformative consent even though nobody can ever know that she has done so. An example like this might seem to provide a simple reductio of mentalism: even if such a person were to mentally accept another party’s “boundary crossing”, it would still surely be wrong for that person to attempt sexual contact with her? But mentalists argue that such an intuition is mistaken. Mentalists have drawn an analogy with the legal notion of abandonment where one’s change of mind can make a normative difference without being communicated (Dsouza 2013; Alexander 2014). For example, suppose R has left her bag at the airport. She cannot be bothered to return to get it so decides to abandon it. Although R’s mental abandonment has not been communicated to anyone, someone who takes the bag—after the decision to abandon has been made—does not wrong the R. In the same way, the mentalist argues, it is what one has in mind, with regards to one’s rights against, say, bodily touching, that makes the moral difference. Here, normative change tracks the intentions and decisions, not facts about whether those mental changes are communicated to others.

But this analogy is weak. Abandonment is more like abrogation—the foregoing or deactivating of a right in general. The change effected by abrogation is a broad, nonspecific one. But consent is not abrogation. Consent has a specific intentional content and a normative scope. The purpose and function of consent is to make a normative difference to the status of some actions for some agents (at a certain time, for a certain purpose, and so on). Consent changes other, specific, agents’ reasons for acting (or for refraining from acting) in certain specific ways. As Wertheimer puts it: ‘If we ask what could change A’s reasons for action, the answer must be that B performs some [explicit behavioural] token of consent. It is hard to see how B’s mental state—by itself—can do the job’ (Wertheimer 2003: 146).⁸

But the mentalist has two responses here (Alexander 2014). First, we need to keep apart two different normative questions: first does S breach the norm in question? Is S culpable? In the moral case, this comes down to the contrast between S wronging R, and S being blameworthy. Culpability and blameworthiness exhibit a distinctive kind of epistemic dependency. Such statuses are determined by what S reasonably believes about R’s consent. In contrast, facts about whether or not a norm is breached, or whether or not S wrongs R are fixed by R’s mental acts alone. Suppose S has no evidence that R has consented to his doing X, S is blameworthy if he does X, even if in fact R has consented.

Second, in response to the objection that the function of consent is essentially social, and that it changes other parties’ reasons for acting, the mentalist can argue that we need to keep apart practical but non-normative aspects of consent transactions, from the properly normative aspects of those transactions. Consider a patient consenting to treatment: it is prudential that she communicate her (mental act of) consent if she wants to surgeon to know that she has consented, but the thing that does the normative work, is her decision to forego her objection, not its communication.

The argument for mentalism thus has two strands. First, there are the considerations—outlined above—in favour of holding that a certain kind of mental act is necessary for normatively transformative consent. Second, the mentalist argues that we do not need to make an appeal to anything else. The mentalist does not deny that consent typically will have a “public face”, nor do they deny that public acts are of great importance in consent transactions. But this is because, the mentalist suggests, overt behaviour plays an epistemic role: that of signifying to others that the appropriate act has been made. Consent behaviour is of epistemic and practical importance because it allows relevant others to know of the consentor’s mental acts, but epistemic and practical importance is not the same thing as normative efficacy.

In sum, certain direct and intuitive lines of objection against mentalism are not conclusive. It does not follow from this that mentalism is correct, only that the arguments deployed against it so far are inconclusive. In the remainder of this paper a stronger case against mentalism is developed.

3 Mentalism and “deceptive” acts of consent

Mentalism has to accept—as a brute matter of fact—that normatively transformative consent is typically effected via overt behaviour, verbal or otherwise. Because the mentalist holds that certain mental acts are sufficient for normative change the mentalist is committed to:

• COMPLETENESS THESIS: the change in normative status distinctive of consent—rendering another’s action morally permissible—is completed by an act within the consentor’s mind.

On the mentalist theory, normative change is brought about in the mind and overt behaviour plays an epistemic and practical role. It allows the consentor to make known her private mental act to others and this can be of considerable practical importance. Suppose R wants to receive surgical treatment but knows that surgery is prohibited without consent. Surgeon S will not treat her unless S has reason to believe that R has changed the normative situation. R’s rights of bodily integrity pose an obstacle to her own treatment, so it is prudential to communicate her mental act of consent to S.

Mental acts are not self-broadcasting: a mental act of consent need not be made known. Relatedly, we can engage in the kind of behaviour that would be taken by others to be an expression of consent—including paradigmatic consent behaviour like saying “I consent”—without having made the appropriate mental act. Mentalism is thus committed to:

• INDEPENDENCE THESIS: there is a logical independence between consent behaviour and the normatively efficacious acts of consent. The behavioural act could be made without the accompanying (mental) normatively efficacious act, and the normatively transformative mental act could be performed without the behavioural act.

If the completeness and independence theses are correct, it should be possible for an agent to engage in an insincere acts of consent where the overt behaviour does not correspond to, or reflect, the underlying mental act. For example:

• BOSS: R has just bought a vintage Triumph Bonneville motorcycle. She believes it to be non-working. R’s boss S hears that R has a new motorcycle (and, having no evidence otherwise, believes that it is fully functional). S asks if she may come round to take it for a ride. If the bike were working R would not permit S to take it for a ride. She does not forego her objection to S’s riding it, because, in her view, it will not be ridden: there is no need to forego her objection. However, in order to gain favour with S, R says “Sure, no problem, why not come round this afternoon and take it for a spin” (planning to feign surprise when the bike does not start). Unbeknownst to R, the Bonneville is in working order. S visits, starts the vintage bike, and takes it for a ride.

Similarly, if the independence thesis is correct, an agent should also be able to engage in insincere refusals.⁹

• SUITOR: R very much wants suitor S to kiss her hand. She has decided in her mind to forego any objection to S’s doing so. S approaches R and asks to kiss her hand. R then notices that they are being watched by another, much wealthier, suitor T. R does not want T to lose interest, and, though she still foregoes her objection to S kissing her hand, she does not want T to know this, so she explicitly says to S, in front of T “No, you may not”. She intends to reveal the truth about her mental act of foregoing her objection to S shortly, when T is out of earshot. However, S impetuously takes R’s hand and kisses it anyway.

Let us assume that in each example R makes her utterance freely. She is not coerced into her utterance, nor is her utterance misdirected by deception. Also, let us suppose, R makes her utterance with the intention that others (falsely) believe that she intends to change the normative situation (as in BOSS) or that she intends that others (falsely) believe that she intends to maintain the normative status quo (as in SUITOR). If R freely and knowingly acts in a way that aims at bringing it about that others believe that she intends to permit S to do X, it is hard to see how S wrongs R, if S acts in accordance with her freely offered act of consent. If R makes an explicit refusal and S goes against it, S surely wrongs R, even if, in R’s mind, she has foregone her objection. This is not mere blameworthiness. To go against someone’s explicit refusal wrongs them, and fails to respect them as another agent with the power to determine whether or not it is permissible for another to touch her.

Mentalism, however, gives us the opposite answer. In BOSS, R has not mentally waived her rights, she is only deliberately and freely acting in a way that seeks to induce the false belief that she consents to the taking of the bike. In SUITOR R does forego her objection in her mind, but decides to induce the false belief that she refuses to permit S to kiss her. In these examples we have two candidate acts of consent (or refusal) in play. We have R’s explicit act, and her mental act. The freely made, intentional, explicit act seems to be a better candidate for normative effectiveness than the purely private act.

Many people may share the anti-mentalist intuition here. But the mentalist may not. Indeed, a reviewer of this paper argues that the mentalist can deal with both SUITOR and BOSS. First, in SUITOR if we assume mentalism, then R does consent, but S does not know this. S is blameworthy, but the facts of the matter are such that S does not in fact wrong R. So SUITOR by itself does not show mentalism to be false. Second, in in BOSS the mentalist can agree with the intuitive judgement that R does permit S to take the motorbike whilst insisting that this is consistent with mentalism. R does consent, but consents conditionally: R consents to S’s taking the motorbike if (but only if) it is working. This is no different, so the objection goes, to someone who bets on a racehorse on the (false) assurance that it will win. If it loses, R has no complaint, because R took a gamble, and this is the situation in BOSS.

Part of the mentalist objection to BOSS is that R actually does consent in her mind, and that is why our intuitions are that the insincere performance changes the normative situation. The idea of a conditional consent is meant to achieve this. But it is not at all clear that the consent given in BOSS is conditional, especially if we fix upon the consentor’s intentions. We can bolster up the example a little. Suppose R has herself disabled the motorbike precisely in order to execute her insincere “pandering” act of consent. It would be very odd to characterize her intention (on the mentalist view) as “I set aside my objection to the motorbike being used if it is working”. This would be a rationally odd thing to entertain. Indeed, the pandering strategy used by R is based on the assumption that the motorbike is not working.

But doesn’t this make it just like a gamble? However, if we take the analogy with gambling seriously this seems to underscore the anti-mentalist intuitions rather than support mentalism. Suppose R makes an insincere bet, we don’t hold that she is thereby exempt from the obligations incurred when the bet is lost. On the flipside, suppose R bets on a horse mentally but without the public commitment. We don’t hold that the normative situation is changed by doing so. Normative change tracks the performance, not what the agent has in mind, and, importantly, a mental bet is not yet a bet.

Finally, it is important to stress that the substantive case against mentalism is not solely grounded in our intuitions about these two examples. Rather than letting intuitions do the work, our case against mentalism will be stronger if we can offer an alternative, independently justified, account of the nature of consent which shows why mentalism is wrong and also explains why it is R’s overt action in BOSS that is normatively transformative.

4 Performatives and robust reason-changing

One alternative to mentalism is that consent is a performative (Wertheimer 1999; den Hertogh 2011). On this view, consent is but one member of a wide family of normatively transformative performatives like promises, vows, prohibitions, and commands, where an agent can change the normative situation just by making the appropriate kind of speech act in particular contexts (providing further “felicity” conditions are met).¹⁰

A performative conception of consent has some considerations in its favour, but also some against it. In its favour, some acts of consent readily pass a simple test for performativity. A test for whether a verb is a performative one—that is, that it has at least some performative uses—is that it can, in some uses, be prefixed with ‘hereby’ (Austin 1962: 57). The ‘hereby’ draws attention to the performative nature of the utterance. “I hereby F” draws attention to the fact that the speaker is intending, by her utterance, to bring it about that she Fs, in virtue of its being recognised as intending to do so (Searle 1989). Consent readily passes the ‘hereby’ test.¹¹ So, in at least some cases—when prefixed with hereby—consent is a performative.

But a performative theory of consent seems to be very limited in its scope. The bulk of everyday consent transactions do not use “hereby” and very many everyday consent transactions do not explicitly use a consent verb at all. Alexander sees such a fact as ruling out the performative theory: he argues that consent cannot be a performative because ‘there is no canonical form of words or action that count as consent’ (Alexander 2014: 2).

One problem with a performative theory of consent is that it saddles itself with a concept—performative—whose roots lie in philosophy of language. This ties the concept to the interests of philosophers of language. Searle (1989), for example, claims that the notion of performative is restricted to those illocutionary acts that are “performed by uttering a sentence containing an expression that names the type of speech act, as in for example, “I order you to leave the room.” (p. 537). Searle may delineate a category of utterances of philosophical interest, but on Searle’s conception of performative, consent would rarely be a performative. But our interests in formulating a theory of consent are not the same as those of philosophers of language with an interest in a taxonomy of speech act verbs. Our interest is in how agents change the normative situation by their acts, and with the question of what kind of act it is that changes the situation in that way.

Consider Searle’s example of an order. An order changes the normative situation by imposing, or generating, an obligation on the ordered party. There are different ways that an order can be effected. Suppose one agent R, has the authority to order another S. She can make the order in a variety of ways. For example:

• (i) “I hereby, by the power invested in me by the King, order you to leave”

• (ii) “I hereby order you to leave”

• (iii) “I order you to leave”

• (iv) “Leave! Now!”

• (v) R waves her hand in a dismissive gesture, intending to order S to leave the room.

Only (i)–(iii) are performatives in Searle’s sense, but our concern is with a broader category—that of acts which bring about changes in the normative situation. This broader category includes all of (i)–(v) above.

Let us stick with orders (or commands) for a moment.¹² A command gives the commanded party a reason to act: commands are a species of reason-giving act. But there are different ways that one person can bring about a change to another party’s reasons for acting (or for refraining from acting). Suppose S and R are walking together. S is holding up an umbrella. R tells S that the rain has stopped. S now has a reason to take down her umbrella, but not because R has given S reasons. R’s act merely plays the role of “triggering”—in Enoch’s words—a reason that S already has. Given that S has reasons (independently of R’s saying so) to let down her umbrella if the rain has stopped, R’s act of speech simply triggers this reason by making it known.

In contrast, performatives like commands involve what Enoch calls robust reason giving. Although robust reason giving is, at root, a species of triggering reason, it has some distinctive features. First, robust reason giving is essentially intentional. Suppose S tells T “It is has stopped raining”, but is (unbeknownst to S) overheard by R. S does not intend to give R a reason, but does so inadvertently. Robust reason-giving essentially involves the intention to give another party reasons. If Colonel R, who has the authority to order troops to attack, unwittingly presses “send” on a prepared message “attack now”, no order is given, though one will seem to be given (and the troops who act upon it are not blameworthy).

Merely having the intention to give another person reasons to act, and getting them act as one intends, is not sufficient for robust reason giving. To adapt one of Enoch’s examples: suppose R is the son of a military dictator, but not a military officer himself: he has no formal authority over soldier S. Suppose R “orders” S to “attack the homes and villages”. Suppose S rightly fears what R’s father might do if S does not comply. Here R intends to give S a reason to attack the villages, but R has no power—has no military authority—to directly give S a duty. R is simply trading upon other reasons that S has (to avoid the ire of the dictator). This is really a species of threat, not command.

In contrast, when Colonel R, with authority over S, orders S to attack, R intends to put S under a duty by making some kind of act—e.g., saying “Attack the homes and villages”—and R intends that S be under that duty just in virtue of S’s recognising that R’s utterance is intended to have this effect. Orders and commands involve a distinctive kind of reflexive and recognition-directed intention.

• (i) R intends to give S a reason to φ, and R communicates this intention to S;

• (ii) R intends S to recognize this intention;

• (iii) R intends S’s given reason to φ to depend in an appropriate way on S’s recognition of R’s communicated intention to give S a reason to φ (Enoch 2011: 13).

Enoch’s focus is upon explicit commands, and it is for that reason that he frames things in terms of “communicating” the relevant intention. But commands, as we have noted, can be given by a wave of the hand, or a nod. This fact points us towards two distinctive features of robust reason-giving.

First, the act by which the reason is given is in a broad sense arbitrary relative to the kinds of reason that are given. Let us clarify this. Consider a non-robust “triggering” reason. Suppose R steps out in front of S’s motorbike, R’s action gives S reasons to brake. R’s act triggers a background reason (e.g., a reason to refrain from injuring others on the road). The reasons that are triggered are very tightly tied to the kind of act that R performs. In contrast, when R orders S to do something, there is no intrinsic connection between the act by which the order is effected—speaking, waving, nodding, writing, shouting—and the reasons that are given. This reflects a kind of arbitrariness that is present in communication more generally. For example, suppose Colonel R has agreed that the order to attack will be given by a coded phrase to deceive any enemy spies. R says “The sun is out in Florida”. What matters is that when R makes this utterance—what makes it a command—is that in that context R performs an act aimed at fulfilling distinctive reflexive, recognition-directed intention outlined above. The arbitrariness here is not absolute. The command has a content and is issued in a context. The act performed has to be of a type that is appropriate to fulfil the intention to put others under an order to do X (rather than Y, say) in that communicative context. But R will know that there are many different ways of doing this, and, which one is chosen will depend upon other factors (time, formality, pragmatic elements, what R takes S to already know; epistemic elements (what will S be able to recognise), and so on).

The second point about performatives of this kind, whether verbal or nonverbal, is that part of the commanding agent’s reasons for engaging in that act is to aim to change the normative situation. If we ask why R says “Attack!” or “The Sun it out in Florida” part of R’s reason for doing so is precisely to change the normative situation. R, of course, may have other reasons (to impress a senior, to end the conflict, and so on). In contrast in our “crossing the road” triggering-reason type of case, R may trigger S’s reason unintentionally (indeed, in the example, R is unaware of S’s presence), R is not crossing the road in order to trigger a reason. This is linked to the relatively arbitrary nature of robust-reason giving acts. If the aim is to fulfil an intention to change the normative situation via others’ recognition of that intention, the type of act one performs is context-sensitive, and, in that context, there may be few other reasons to perform such a relatively arbitrary act other than to fulfil one’s normative intentions.¹³

5 Objections and replies

At this point a mentalist about consent may wonder what relevance our discussion is to consent. The mentalist may argue that commands put another party under an obligation, whilst consent involves the setting aside of an obligation. It is because they put others under an obligation that commands have to be communicated, but consent does not.

But now consider the fact that in many contexts, commands are revocable. Suppose R issues the command “Attack the villages” but then, a few minutes later (on the basis of a new strategy, say) says “Stop! Cease the attack”. The question now is: how does this utterance change subordinate S’s reasons? It would be odd if the command were robust, but the revocation were not. When R revokes the order, R intends to change S’s reasons via an act that aims at fulfilling a variant on the distinctive reflexive and recognition-dependent intention noted earlier.

• (i) R intends to change S’s reason to φ,

• (ii) R intends S to recognize this intention;

• (iii) R intends the change in S’s reason to φ to depend in an appropriate way on S’s recognition of R’s intention to change S’s reason to φ.

• (iv) R performs an act X which aims to fulfil the intentions in (i)–(iii).

This broader schema covers not just robust reason-giving, but robust reason-changing. The revocation of a command involves R setting aside an obligation that S has via the performance of the appropriate act with the relevant intention. But, structurally, this is exactly what goes on with consent.¹⁴ By the act of consent, R intends to set aside S’s reasons to refrain from acting in certain ways via S’s recognition that R intends to do so.

The mentalist might now object that consent is not an authoritative form of reason-changing and is thus not robust. We can concede that the kind of authority that people have in giving consent is not the same as the institutional authority that a colonel has, say. But the notion of authority that matters here is that which is tied to normative powers, not to this or that specific institutional implementation or grounding of those powers. The reason-changing that matters to us is that, in consent, R has the power, just in virtue of performing the appropriate kind of act (and it need not be a verbal act, let alone one that deploys the word “consent” or “I hereby consent”) with the requisite intentions (and no defeating conditions) that other parties reasons are thereby changed, just because the person has intended to do so. There are of course further questions as to why the individual should have such an authority to change the normative situation, but our focus is on the way that the normative situation is changed, not in its fundamental grounding. That consent has a distinctive kind of authority can be illustrated by a fictitious example where it fails to be respected:

• CONSENT WITHOUT AUTHORITY: R’s surgeon S proposes surgery to R, seeking her consent. R explicitly gives her permission (via a signed consent form). S then says “Thank you for that, I shall take that information into account in my deliberation about whether it is permissible to treat you.”

This example should seem odd. This is because (medical) consent transactions involve a distinctive kind of respect for the consenting party.¹⁵ When a competent adult patient consents or refuses, that is, as it were, the last word. The patient has the final say about whether to create an exception to her rights of bodily integrity. This kind of respect is a reflection of the authority of consent, something that is analogous to the authority to revoke a command.

The mentalist might now object that although consent is authoritative in some sense, the reason why we ought to take the consentor’s word as final is because of her first-person epistemic authority with regard to her mental acts. This would maintain a commitment to the completeness and independence theses noted earlier. But such an objection fails to engage with the robust reason-changing nature of consent. An act of consent involves the consentor acting in a way that aims at fulfilling a recognition-directed intention. But how can an agent seek, by doing some act X, to fulfil her intention that another party change her reasons via that other party’s recognition of that act, if the act in question is entirely mental? How can R intend that S recognise act X as having some kind of status if R does not do anything that would put S in a position to recognise it? It is no good here for a mentalist to argue that consent needs to be communicated for prudential or practical reasons. The point here is that the agent cannot complete an act of consent in her mind at all. She can have the intentions that another change her reasons, but unless and until she acts in such a way that is aimed at another party recognising such acts as fulfilling the appropriate intentions, an act of consent has not been made. The robust reason-changing nature of acts of consent is at odds with both the completeness and independence theses.

The mentalist might object at this point that the robust reason-changing theory mistakenly ties normatively transformative acts of consent to their being recognised by others, or their being known by others. But that would be to misrepresent the theory offered here. The robust reason-changing account holds that there is a difference between performing a normatively transformative act of consent and others knowing of it. R can permit S to do X without S ever knowing that she has done so. But the robust reason-changing account accommodates this fact readily. For example, suppose R leaves a note permitting S to borrow her motorbike “Feel free to take the Bonneville for a spin”. Suppose S never sees the note. The act of consent—the change in the normative situation—is made when R performs the appropriate act with the aim of fulfilling the appropriate intention, whether or not S comes to know of that act, is another thing. If S takes the bike without reading the note, she is blameworthy, but does not wrong R. None of this implies that consent—or command for that matter—is the kind of act that can be performed within the mind alone.

The robust reason-changing account holds that acts of consent—like other robust-reason changing acts—are aimed at fulfilling a reflexive and other-directed intention. But this cannot be done in the mind alone. A command does not serve to make known some already completed mental act of command. The behavioural act is not something that plays a merely epistemic role: such behaviour is constitutive of the act of command. By analogy, consider a promise. A promise does not serve to make known some already completed promise made in the mind, it aims at effecting the promise. A command does not seek to make known some already completed command in the mind. The same is true of consent. Command, promise and consent are constitutively aimed at other people’s recognition of those acts as ones which are performed with a distinctive intention to change the normative situation. One cannot coherently perform such an act whilst aiming to complete it in the mind alone. This is not to claim that an act of consent is only made, or only changes others’ reasons, when it is actually recognised by relevant others. The problem for mentalism arises because of the kind of intention that robust reason-changing involves, not because it collapses the distinction between acts of consent and their recognition.

But the mentalist can object at this point that we are conflating different kinds of reason. A mental act of consent changes other agents’ objective reasons (it sets aside a reason they have to refrain from acting in certain ways), but unless this change in objective reasons is made known to others, there is no change to the subjective reasons that others have. Thus, in the “abandonment” example noted earlier, if R has set aside her rights in her mind, objective reasons have changed, even if nobody’s subjective reasons have done so.

The first response to this is that this point seems to re-iterate the distinction between wronging and blameworthiness that we have already discussed. The second, more important, response, is that this objection may seem to have some plausibility if we view acts of consent as a kind of Platonic moral bookkeeping exercise, where an agent, in her mind alone, brings about a change in “normative score” without the act of consent being aimed at being recognised by others as an act of that kind. Now, in some cases what we think, or decide, might make a difference to the moral status of others’ actions even if no behavioural act is performed. For example, suppose we hold that R is wronged by S if R says something deeply offensive to R’s religious convictions. Now suppose her convictions change but nobody else knows. An act that would have wronged her, will now no longer do so, and this is because her state of mind has changed. But this kind of mental change is not the exercise of a normative power, and does not involve anything like robust reason-changing. In a similar way, simply deciding to forego one’s objection is not yet to have performed an act of consent, just as merely deciding to order a subordinate, or deciding to promise, does not (yet) give an order or make a promise.

If we do adopt a distinction between subjective and objective reasons, as the objection above suggests, there is still a serious problem for mentalism when we consider the way that acts of consent change others’ reasons robustly. When R consents to S’s actions, she is not merely aiming to change his subjective reasons, she is aiming to change his objective reasons. Suppose R sets aside her objection in her mind to S’s taking her bike. On the mentalist proposal this mental act by itself changes the objective reasons that S has, even if it does not yet change S’s subjective reasons. But what kind of change is this? It looks like it has to be an instance of merely triggering change, rather than robust reason-changing. Why? Because robust reason-changing involves the distinctive kind of other-directed, reflexive, intention noted above. It is unclear how simply setting aside one’s objection in one’s mind can be an act aimed at fulfilling that kind of intention. In order to sustain the objection, the mentalist would have to offer an alternative account of robust reason-changing, or explain why acts of consent (contrary to all appearances) do not change reasons robustly (they would also then have to explain the disanalogy between consent and other robust-reason-changing performatives like promise and command, or be saddled with the serious challenge of giving a mentalist account of those performatives too). The distinction between objective and subjective reasons is independent of, and orthogonal to, the distinction between acts which must be aimed at fulfilling other-directed intentions (that is, robust-reason changing acts), and those that are not (a latter set which would include triggering reason-changes).

Note that the central arguments offered here against mentalism also pertain to certain variants of the so-called “hybrid” theory of consent (Malm 1996). The hybrid theorist holds that a mental act alone is not sufficient for consent, but retains a commitment to the idea that public acts of consent signal or express consent. But robust reason-changing is not a species of expressing or signalling some normatively salient act. A public act of consent—be it verbal, or nonverbal, be it explicit, or explicitly prefixed with “I hereby”—does not signal a private act of consent, it effects the act of consent, just as R’s order to attack does not signal a private mental act of command.

6 Conclusion

We have been critically assessing mentalism: the view that normatively transformative acts of consent can be effected within the mind alone. Initial objections to mentalism can be met by the mentalist with relative ease, especially if we accept the assumptions that it makes sense to think of an act of consent as something that can be completed in the mind, and assume that the role of explicit behaviour is to communicate, or make known, such an act. With these assumptions in place mentalism cannot be rebutted by the observation that consent is social, directed at others, or gives reasons to others. The mentalist can argue that these latter elements are secondary, practical, elements, which may make a difference to blameworthiness, but are not essential to normatively transformative consent. Acts of consent are one thing, their communication and reason-giving nature another.

The examples of “deceptive” consent raised a question mark about this kind of separation between consent and behaviour. These examples were not meant to provide a conclusive case against mentalism, but, rather, were introduced to motivate an explanation of our anti-mentalist intuitions in these examples. We now have that explanation. In BOSS, R performs a robust reason-changing act, with the appropriate reflexive, recognition-directed intention. She intends by her utterance that S recognise this utterance and intending to permit S to ride the bike. But R does not actually forego her objection, she (falsely) believes that the bike cannot be ridden so there will be no objection to forego. In the case of SUITOR, the fact that R has consented—foregone her objection—to S’s kiss in her mind is not yet to exercise her normative power. She has done nothing that aims at changing the normative situation via an act aimed at fulfilling the appropriate kind of reflexive, recognition-directed, intention, indeed, she has made explicit her refusal and, given the authority of her normative power, that ought to be the last word.

The mentalist cannot respond that we have conflated wronging with blameworthiness, for our robust reason-changing account accepts that there is a difference between performing a normatively transformative act of consent and its being known by others (e.g., our example above of the unread note consenting to the use of the motorbike). Nor can the mentalist respond that we have conflated subjective and objective reasons, because the problem that arises for the mentalist theory is to account for robust-reason changing, which, in turn, requires acts aimed at fulfilling other-directed intentions.

The robust reason-changing theory of consent explains how it is that “consent” can be used in explicit performatives and why it can be prefixed, in some contexts, with “hereby”. On the mentalist account, it is very odd why anyone would say that they hereby consent—why are they not simply reporting their consent, rather than effecting it? But the advantage of our robust-reason changing account is that it does not need to saddle itself with the “performative” label. On our account normatively transformative acts of consent do not need to be effected with a consent verb (in whatever language), indeed, they do not need an act of speech at all. Our theory can thus readily accommodate the fact that consent involves a very wide range of behaviours, verbal and nonverbal. This meets the mentalist objection that ‘there is no canonical form of words or action that count as consent’ (Alexander 2014: 2); the same is true for commands and many other robust-reason changing acts.

In conclusion, our critical rejection of mentalism has allowed us to develop a clearer, and more defensible, conception of the fundamental nature of permissive consent (what is sometimes referred to as the ontology of consent). The robust reason-changing account does not readily fit with extant positions on consent, but shares a good deal with the “performative” theory. The robust reason-changing account raises plenty of questions, about why and when it is that we have this kind of robust reason-changing power (and how that power varies in different normative contexts). It raises questions about how the power exercised—e.g., about how the normative scope of an act of consent is fixed—and about how the exercise of such a power might be undermined (e.g., by deception). These are matters for another occasion. If we are to make sense of consent it is important to start with a correct conception of what consent is. Consent is not a mental act, nor is it the communication of a mental act, it is a behavioural act aimed at fulfilling a distinctive, reflexive and recognition-directed intention.