Consenting Under Third-Party Coercion


This paper focuses on consent and third-party coercion, viz. cases in which a person consents to another person performing a certain act because a third party coerced her into doing so. I argue that, in these cases, the validity of consent depends on the behavior of the recipient of consent rather than the third party’s coercion taken separately, and I will specify the conditions under which consent is invalid. My view, which is a novel version of what I call a Recipient-Focus-View, holds that coercion invalidates consent only if consent was ‘obtained by’ coercion, but not if consent was ‘merely motivated by’ coercion. I explain and support my view on the basis that it best reconciles an unnoticed tension between two fundamental principles in the debate on consent (which I call the Coercion Principle and the Permissibility Principle) and that it can deal with cases that undermine other Recipient-Focus-Views.


Introduction
When we consent to a medical procedure, sex, or a tattoo, we permit people to act in ways that would otherwise be impermissible. But our consent is valid, i.e. it creates a permission, only if we have not been coerced into giving it. Relying on this common sense description of consent, one may extract the following two principles fundamental to the ethics of consent: twisting that arm is necessary to save many people's lives (wide permissibility) (Nagel, 1986, p. 176).
In this paper, I am concerned with narrow permissibility only, as this is the kind of permissibility that people create when they give valid consent. This distinction between narrow and wide permissibility is widespread in the ethics literature and the debate on consent, even though often expressed with different terminology (Dougherty, 2015, p. 232;Mayr, 2019, pp. 118-124;Schaber, 2018, p. 56;Tadros, 2016, p. 201;Thompson, 2004, pp. 333-335).1 Moreover, I will assume that the Permissibility Principle is consistent with the view that an act which normally requires valid consent, e.g. grabbing a person's arm, may sometimes not require consent at all, e.g. when grabbing that person's arm is part of justified self-defense against the person. This is because, in such cases, the person forfeited a right not to be grabbed, and thereby made consent redundant. Thus, the Permissibility Principle operates on the assumption that the relevant rights (e.g. not to be touched or treated in certain ways) have not been forfeited. Since the forfeiture of rights is a discussion separate from consent, and since my examples about tattoos, sex, and medical procedures will not present cases forfeiting consent-sensitive rights, I will not discuss forfeiture in this paper.
On the other hand, the Coercion Principle holds that coercion invalidates consent. More precisely, it holds that coercion always invalidates consent, irrespective of who exerts it; and for the purpose of this paper, I will adopt the canonical understanding of coercion as the use of illegitimate, severe threats to make a person comply with a specific demand (see Kiener, 2020). In such cases, inflicting the threatened harm would also violate an important moral right, such as the right to bodily integrity.2 The Coercion Principle is also widely held (Appelbaum, Lidz, & Klitzman, 2009, p. 33;Feinberg, 1986, p. 196;Ferzan, 2018Ferzan, , pp. 1000Ferzan, -1001Millum, 2014, p. 114;Nelson et al., 2011, p. 8). Millum, for instance, supports the Coercion Principle as follows: he argues that "valid consent requires that the person who has the authority to make a decision make that decision herself" (Millum, 2014, p. 118). If someone is under an illegitimate controlling influence, e.g. coercion, she is "not making the decision ( … ) herself" (Millum, 2014, p. 118) and hence her consent is invalid, no matter who exerted the coercion. Finally, the Coercion Principle can be found in most European jurisdictions in civil and criminal law.

Inconsistency in Cases of Third-Party Coercion
Inconsistency between the two principles arises once we consider a case of third-party coercion, i.e. a case where a person is coerced into consenting not by the recipient of consent but by a third party. I will assume that someone is a 'third party' if and only if he is not the recipient of consent and does not collaborate with the recipient. The following case, taken from a paper by Joseph Millum, provides a good example of third-party coercion: Tattoo "Coercer threatens to beat up Victim unless she gets a tattoo from Recipient's tattoo parlor. Recipient has no particular interest in tattooing her, but he cannot otherwise prevent the threat from being carried out and, although she would otherwise not want a tattoo, the beating is clearly worse for Victim than the tattoo. When she explains the situation to Recipient, Victim confirms that she prefers that he tattoo her than take any other action" (Millum, 2014, p. 118).
Coercer is a third party: he is not the person receiving Victim's consent and does not collaborate with Recipient. Yet, his threats clearly coerce Victim. Recipient is the person receiving consent and, let us assume, Recipient behaves flawlessly towards Victim. Recipient would offer anything he could to help Victim, but Recipient cannot help Victim in any way other than executing the tattoo and Victim requests the tattoo for this reason. If Recipient executes the tattoo in this case, he does not take advantage of the situation and respects Victim's autonomy as much as possible. Hence, Millum argues, and I agree, that Recipient permissibly executes the tattoo. Millum made an important point: Tattoo shows that one may still permissibly tattoo another person even if that person has been subjected to third-party coercion. Unfortunately, Millum stops here. He does not use Tattoo to show, as I think he could, that there is a fundamental conflict between the Permissibility Principle and the Coercion Principle. Thus, to outline this conflict, I need to extend the analysis of Tattoo beyond what Millum himself says and, in particular, determine whether Recipient's execution of the tattoo is narrowly permissible or merely widely permissible. This is because the Permissibility Principle will only be triggered if Recipient in fact acts narrowly permissibly when executing the tattoo. Recipient acts narrowly permissibly if and only if Recipient's execution of the tattoo does not wrong Victim, i.e. violate Victim's right to bodily integrity. In such a version of Tattoo, and absent other reasons against tattooing, Recipient would also not engage in any wronging at all when executing the tattoo, despite the third party's coercion.3 By contrast, Recipient acts merely widely permissibly if Recipient's execution of the tattoo is justified, all things considered, but still wrongs Victim. On this understanding, executing the tattoo is widely permissible because sending Victim back to the mercy of Coercer would be even worse and wrong Victim even more seriously. But executing the tattoo still wrongs Victim (i.e. is narrowly impermissible) because it violates Victim's rights to bodily integrity, given the coercion in the background. Accordingly, Recipient acts permissibly, not because Recipient does not wrong Victim, but because Recipient chooses the lesser of two evils.
So, which is the correct view? I claim that it is the first one, namely that Recipient's execution of the tattoo is narrowly permissible, for two reasons. Firstly, none of the features which accompany narrow impermissibility, i.e. wronging, are present in the interaction between Recipient and Victim. If a person is wronged by someone else, that person is entitled to demand an apology from the wrongdoer, and may resent and potentially forgive him (Bennett, 2018, p. 209 ff.;Darwall, 2006, chapter 5;Owens, 2012, chapter 2;Wallace, 2014, pp. 119-142). The wrongdoer, on the other hand, incurs -at a minimum -an obligation to apologize. However, in Tattoo, Victim cannot justifiably resent Recipient for executing the tattoo, demand an apology, or offer to forgive Recipient. Moreover, Recipient is not obliged to apologize to Victim. After all, Recipient respected Victim's autonomy as much as he could and Victim requested the tattoo. Compare the interaction between Recipient and Victim with the interaction between Coercer and Victim, where it is clear that Coercer's threat wronged Victim. Victim can indeed justifiably resent Coercer, demand an apology, and perhaps eventually forgive him; and Coercer is certainly obliged to apologize. Therefore, it seems that only Coercer but not Recipient wrongs Victim. Recipient's execution of the tattoo is narrowly permissible (and not just merely widely permissible).
Secondly, the oppositite view to narrow permissibility, viz. considering the execution of the tattoo to be merely widely permissible, is either (i) intensionally inadequate or (ii) collapses into the narrow permissibility view.
(i) On a mere wide permissibility view, Recipient acts permissibly because he chooses the lesser of two evils. On a standard interepretation of such a view, something is the lesser evil when it imposes the lesser harm.4 Thus, so understood, executing the tattoo is justified because it imposes the lesser harm on Victim, compared to the alternative of sending Victim back to Coercer, who would severely hurt Victim.
But such a comparative assessment of harms also extends to cases in which Victim does not request the tattoo or even rejects it. Even when Victim rejects the tattoo, tattooing Victim may be the lesser harm. Admittedly, if Victim rejects the tattoo, executing the tattoo may nevertheless impose additional psychological harm on Victim. However, as long as Coercer would inflict even greater harm, Recipient can still be justified in tattooing Victim against her request on the mere wide permissibility view.
But the extension to such cases is mistaken. The permissibility of Recipient's conduct in Tattoo does not depend on a comparative assessment of harms and is not independent of Victim's request for the tattoo in this way. The permissibility of tattooing requires Victim's request and that Recipient respects the residual control that Victim has left in the constrained circumstances. Thus, this version of the mere wide permissibility view is intensionally inadequate, viz. it falsely suggests that Recipient acts permissibly because tattooing imposes the lesser harm, rather than because Recipient respects the residual control that Victim has left in the constrained circumstances.
(ii) For this reason, one may opt for a different version of the mere wide permissibility view and claim that the lesser 'evil' does not necessarily mean the lesser harm. Alternatively, it could also mean the lesser wrong in the sense that tattooing Victim, when Victim requests it, is less disrespectful than sending Victim back to Coercer, independently of how harmful it is in the end. So understood, the mere wide permissibility view could recognize Victim's request as the central consideration without relinquishing the claim that executing the tattoo still wrongs Victim (due to the fact that Victim was coerced into it).
Although this interpretation is an improvement over the harm-based view, it ultimately collapses into the narrow permissibility view. If Recipient acts according to Victim's request, Recipient ensures what Gerver calls relative autonomy (Gerver, 2021): relative to the interaction with Recipient, Victim enjoys as much autonomy as possible, in the sense that Recipient did not withhold any options from Victim, did not attempt to override Victim's decision on the basis of his own (harm-based) considerations, and only acted as authorized by Victim.5 But if, in executing the tattoo, Recipient ensures such relative autonomy, Recipient can no longer wrong Victim. This is beause wronging (as opposed to general wrongdoing) is the violation of relative claims, i.e. claims that one person (e.g. Victim) has in relation to another person's (e.g. Recipient's) conduct; and the provision of relative autonomy precisely precludes the violation of such relative claims. Thus, as soon as the mere wide permissibility view makes Victim's request central to its approach and thereby shows that Recipient ensures relative autonomy for Victim, it collapses into the narrow permissibility view, the view that Recipient does not wrong Victim anymore.
Hence, I will assume that Recipient's execution of the tattoo is narrowly permissible. On the basis of this extended analysis of Tattoo, I am now in a position to build further on Millum's insights and explain the inconsistency between the Coercion Principle and the Permission Principle. The Coercion Principle yields the following reasoning about Tattoo.
However, once we agree that Recipient acts narrowly permissibly, applying the Permissibility Principle yields the opposite conclusion: 4. Permissibility Principle: Permissibly (narrow) executing a tattoo requires valid consent. 5. Verdict on Tattoo: Recipient acts narrowly permissibly in executing the tattoo. 6. Conclusion: Victim's consent is valid.6 5 Gerver's own formulation of relative autonomy is slightly different. It is presented in terms of options which Recipient is under a duty to provide or not to provide. However, applied to Tattoo, Gerver's view clearly implies that Victim has relative autonomy vis-à-vis Recipient and, on these grounds, also that Recipient does not wrong Victim. Also, Tadros' recent view would, along similar lines, also imply that Recipient does not wrong Victim (Tadros, 2021

2.1
Outline of Millum's Solution Although Millum did not address the dilemma as I set it out, his statements may still allow us to form a potential solution. Before I reconstruct this solution, however, one clarification is in order: the inconsistency between the Coercion Principle and the Permissibility Principle is not a problem specific to Millum's view. It is a problem for any theory about consent. Thus, so far, Millum's view is not part of the problem, but potentially part of a solution. However, to be part of a solution, Millum's view needs to be interpreted -at least for the sake of argument -as holding that Recipient acts narrowly permissibly when executing the tattoo. Otherwise, Millum's view would fail to acknowledge the existing inconsistency between the Coercion Principle and the Permissibility Principle. Hence, I will henceforward attribute the, to my mind, correct narrow permissibility view to Millum too. With this clarification at hand, what is the solution that Millum could provide?
Millum accepts (1), the Coercion Principle: "Victim's consent is always rendered invalid by third party coercion" (Millum, 2014, p. 119). He also accepts (2): Victim is coerced into consenting. Finally, Millum concludes (3), i.e. Victim's consent is invalid. But Millum thinks that Recipient still acts (narrowly) permissibly. Millum claims that "illegitimate third-party coercion does not entail that it is impermissible to proceed with the act consented to even if it means that the consent is invalid" (Millum, 2014, pp. 118-119). Whether or not Recipient may proceed on the basis of invalid consent depends, according to Millum, on "the other options available to Recipient, and what Victim wants Recipient to do" (Millum, 2014, pp. 118-119). Developing these statements, Millum argues that, in cases of coerced and invalid consent, Recipient "may proceed if doing so would constitute a reasonable joint decision of her and the victim of coercion" (Millum, 2014, p. 126). Joint decision-making "involves each party taking the viewpoint of the other as seriously as her own and deciding on this basis what to do" (Millum, 2014, p. 121). If tattooing is based on such joint decision-making, "Recipient thereby respects ( … ) [Victim's] autonomy by giving her as much control over her consent decision as he is able" (Millum, 2014, p. 119). In such cases, "a wrong has been committed by Coercer. But Recipient is not culpable for this wrong: he has done the best he could in non -ideal circumstances" (Millum, 2014, p. 119).7 Hence, Millum claims that Recipient can execute the tattoo permissibly not only if he is given valid consent (option 1) but also if he is given invalid consent but acts on the basis of a reasonable joint decision (option 2). However, it is noteworthy that, in Millum's view, Recipient acts permissibly in Tattoo only if Victim actually requests the tattoo. In Millum's view, a request is still necessary to make a recipient's act permissible; and the absence of a request is sufficient to make Recipient's act impermissible.
Thus, together with the assumption that Recipient executes that tattoo narrowly permissibly, Millum's view can create the following solution to the dilemma: 1. Coercion Principle: Coercion invalidates consent. 2. Facts in Tattoo: Victim is coerced into consenting.

Conclusion: Victim's consent is invalid.
But Millum needs to change the second set of claims: 4.* Modified Permissibility Principle: Permissibly (narrow) executing a tattoo requires either (i) valid consent or (ii) a reasonable joint decision. 5. Verdict in Tattoo: Recipient acts narrowly permissibly in executing the tattoo. 6.* Modified Conclusion: Victim's consent is either valid or invalid.
On the Modified Permissibility Principle, the original conclusion (6), i.e. that consent is valid, no longer follows: consent could be either valid or invalid and, as it happens, consent is invalid in Tattoo and Recipient acts narrowly permissibly due to basing his action on a reasonable joint decision. Hence, there is no contradiction anymore: conclusion (3) and modified conclusion (6*) are consistent with each other.

2.2
Critique of Millum's Solution Unfortunately, this solution reconstructed from Millum's account remains unsatisfactory. Firstly, Millum's overall framework may ask for too much in one respect. Millum claims that, in the presence of third-party coercion, Recipient executes the tattoo permissibly if and only if doing so is based on a reasonable joint decision, and he explains that "[j]oint decision-making involves each party taking the viewpoint of the other as seriously as her own and deciding on this basis what to do" (Millum, 2014, p. 121, emphasis added). Yet, there can be cases where Recipient's execution of the tattoo is not based on a reasonable joint decision, but Recipient may still proceed narrowly permissibly. To see this, consider the following case: Manipulation Someone threatens Recipient with death if he tattoos Victim.
Thus, if Recipient tattoos Victim, he will be killed, and if he does not, Victim will be beaten. To make Recipient proceed, Victim illegitimately manipulates Recipient into tattooing her: she uses emotional triggers and harnesses Recipient's decisional weaknesses.
In Manipulation, tattooing is not based on a reasonable joint decision. After all, Recipient faces greater harm than Victim and, due to Victim's manipulation, it is not true that the decision-making "involves each party taking the viewpoint of the other as seriously as her own and deciding on this basis what to do" (Millum, 2014, p. 121). But even if this is true, it would not follow that, for this reason, it is narrowly impermissible for Recipient to execute the tattoo, i.e. that Recipient commits some wrong when proceeding on that basis. Hence, whether Recipient executes the Tattoo narrowly permissibly does not depend on whether doing so equally accounts for both viewpoints, i.e. whether tattooing is based on a reasonable joint decision. It only depends on whether Victim's viewpoint has been sufficiently accounted for, as it is Victim who will be subjected to the tattoo. Millum's focus on a joint decision fails to capture this important point and, therefore, provides a misleading framework for evaluating when Recipient executes the tattoo narrowly permissibly. In response, Millum could argue that his view only states a sufficient but not necessary condition for the narrow permissibility of Recipient's conduct. But there are two problems with this reply. (i) Millum's account would be incomplete. After all, we still need to know not only what is sufficient, but also what is necessary for Recipient to act narrowly permissibly. (ii) This reply falsely suggests that, when there is a reasonable joint decision, part of the reason why Recipient does not wrong Victim is that Recipient sufficiently accounted for his own interests, for otherwise it would not be a joint decision. But accounting for one's own interests cannot (save in very exceptional cases) be part of the reason as to why one does not wrong others. Not wronging others depends on whether one sufficiently accounted for their interests. Thus, Millum's general focus on joint decisions does not offer an appropriate explanation (or, in other words, is intensionally inadequate), even when it delivers the correct verdict (e.g. no wronging).
Secondly and over and above Millum giving the wrong explanation of why Recipient acts narrowly permissibly in executing the tattoo, Millum's Modified Permissibility Principle also lacks justification for introducing a 'reasonable joint decision' as a concept distinct from valid consent. After all, a reasonable joint decision shares all the essential features of valid consent. Just like valid consent, a reasonable joint decision requires Victim's request for the tattoo, demands that Recipient not control Victim's decision-making, releases Recipient from the obligation not to interfere with Victim's body, and makes executing the tattoo narrowly permissible. Thus, it seems that Millum introduces a reasonable joint decision as something other than valid consent when it actually should be considered an instance of valid consent.
The first point of these criticisms and my example Manipulation particularly target Millum's view in particular and his account of a 'reasonable joint decision.' However, the second point also applies to any other view that preserves the Coercion Principle and modifies the Permissibility Principle in some similar way. By way of explanation, even if other views reject Millum's 'reasonable joint decision' and adopt the correct view on when Recipient executes the tattoo in a narrowly permissible way (e.g. when Victim issues 'request X'), these versions will still lack justification for introducing 'request X' as something distinct from valid consent. After all, such 'request X' will still share all the essential features of valid consent. Just like valid consent, it will require Victim's request for the tattoo, demand that Recipient not control Victim's decision-making, release Recipient from the obligation not to interfere with Victim's body, and make executing the tattoo narrowly permissible. Thus, my critique applies, more generally, to any views that accept the Coercion Principle and aim to modify the Permissiblity Principle. I therefore suggest rejecting Millum's solution alongside other views that also accept the Coercion Principle without modification. In the next section, I will provide an alternative solution that leaves the Permissibility Principle untouched. The Influence-Focus-View holds that whether an influence vitiates consent depends solely on the nature, circumstances, and effect of the influence, but not who exerted it. People defending the Influence-Focus-View still diverge on what they think makes an influence consent-vitiating. Whereas some hold that an influence invalidates consent if it is substantially controlling (Feinberg, 1986;Nelson et al., 2011), others argue that an influence invalidates consent if the influencer lacks the right to exert it (Appelbaum et al., 2009). Yet others hold that an influence needs to be both substantially controlling and illegitimate (Millum, 2014). However, all of them think that whether consent is valid depends only on the particular features of an influence, and not on who exerted it; and all of them also think that coercion is a paradigmatic example of an influence that invalidates consent. Hence, they all accept the Coercion Principle: once an influence qualifies as coercion, it vitiates consent, irrespective of who exerts the coercion.
However, the Influence-Focus-View is not without an alternative. The alternative I want to present in this paper is what I call the Recipient-Focus-View. According to the Recipient-Focus-View, the validity of consent depends on the conduct of the person receiving consent rather than on the third party's coercion alone. Most such views argue, more specifically, that consent is valid if the person receiving consent did not wrong Victim in order to receive consent and invalid if he did wrong Victim in that way. Thus, the focus is on the person receiving consent rather than, as the Influence-Focus-View has it, on whether anyone, including third parties, coerces the person consenting.
As a result, the Recipient-Focus-View rejects the Coercion Principle, according to which coercion always invalidates consent. The Recipient-Focus-View holds that whether coercion invalidates consent depends on the role of the consent-receiver, i.e. Recipient. More specifically, most of these views hold that coercion only invalidates consent if Recipient committed some wronging and, as Tattoo showed, this does not always have to be the case.
But note: on such Recipient-Focus-Views, Recipient's wrong that invalidates consent is distinct from the wrong in acting on invalid consent. The former normally occurs in the process of obtaining consent and prior to Victim's consent, e.g. when Recipient coerces Victim or collaborates with Coercer to extract consent.8 Thus, the Recipient-Focus-View still accepts the traditional explanatory order: the invalidity of consent explains why acting on such consent is wrong, and not the other way around. The Recipient-Focus-View just adds that the reason why consent is invalid in the first place is always due to another, prior wrong committed by Recipient.
But how exactly can Recipient invalidate consent that results from coercion? What kinds of conduct are relevant? I suggest a tri-partite account. Hence, exploiting coercion means that Recipient uses the third party's coercion to force Victim to consent to something which even the coercing third party had not demanded from Victim, e.g. an additional extortionate price (Tattoo Money) or sex (Tattoo Sex). In these cases, Recipient's additional demand is not necessary for Victim to escape Coercer's threat, it imposes greater harm on Victim, and is not one that Recipient is entitled to make.
So understood, exploiting coercion is specific to the particular object of consent, i.e. what Victim consents to. In Tattoo Money and Tattoo Sex, Recipient's exploitation invalidates Victim's consent only to the extortionate price and to sex, but not to the tattoo. This is because exploitation requires a means-end link between Recipient's misconduct and what Victim consents to. And since Recipient exploited the coercion only with regard to the price and sex, but not the tattoo, which Victim would have consented to anyway, it is only Victim's consent to the price and to sex that is invalid.
Moreover, exploitation is more like 'taking advantage' rather than 'getting an advantage,' i.e. benefiting from what Coercer forced Victim to undergo. Exploiting, like taking advantage, requires that Recipient illegitimately harness the third-party coercion to obtain consent to something which even Coercer had not demanded from Victim. Therefore, however much Recipient otherwise benefits from executing the tattoo, such benefits alone do not imply that Recipient also exploits the coercion as I described it.
So understood, my view is consistent with views such as Alan Wertheimer's, according to which there can be instances of mutually advantageous and consensual exploitation. This is because Wertheimer specifically requires the absence of coercion as a necessary condition for such a transaction to be consensual (cf. Wertheimer, 1996, p. 10;2012, p. 283).
In addition, it is noteworthy that Wertheimer also relies on a distinction between different objects of consent similar to the one I presented in Tattoo Sex and Tattoo Money. In an example (called Rescue) in which A charges B an extortionate fee to free B's car, which would otherwise be stuck for hours, Wertheimer says: "Rescue can involve at least two different types of moral transformation. Consider two questions: (1) does B's consent give A permission to tow B's car? (2) Does B's promise (consent) to pay A $200 obligate B to pay the full amount? I believe that the answer to (1) may be yes, because A has not coerced B and because B may be fully informed about the alternatives available to her and competent to choose among them. At the same time, the answer to (2) may be no." (wertheimer, 2010, p. 112) Thus, just as I distinguished between Victim's consent to the tattoo on the one hand and Victim's consent to the extortionate price or sex (in Tattoo Money kiener and Tattoo Sex) on the other, Wertheimer distinguishes between two objects of consent as well.

3.1.3
Facilitating Coercive Success Finally, Recipient wrongs Victim and invalidates consent when he facilitates coercive success, i.e. when he denies Victim an alternative to Recipient's acting on consent (e.g. to executing the tattoo), which would also allow Victim to avoid the threatened harm and not impose an unreasonable cost on Recipient (or others).9 Consider the following example:

Mischievous
As in Tattoo, except Recipient could make Coercer withdraw the threat simply by giving him a phone call. As Recipient knows, Coercer is the biggest fan of his tattoo art and would comply with anything Recipient tells him. Victim asks for help but Recipient refuses. Victim then consents to the tattoo and Recipient excecutes it.
Making Coercer withdraw the threat is an alternative to acting on consent, allows Victim to avoid the threatened harm, and does not impose an unreasonable cost. Hence, Recipient could easily enable Victim to return to the pre-coercion setting, in which Victim could avoid both the threatened harm and what Victim was forced to consent to. Denying Victim this alternative wrongs Victim, facilitates coercive success, and invalidates Victim's subsequent consent to the tattoo.
The key feature in cases like Mischievous, and the reason why I picked the formulation facilitating coercive success, is that Recipient now has control over Victim's situation that normally only Coercer has, i.e. the control over whether Victim can return to the pre-coercion setting and avoid both evils, i.e. the threatened harm and whatever results from complying with Coercer's demand. In other words, it is Recipient's own conduct that now determines whether Victim needs to suffer anything at all. Therefore, if Recipient denies Victim a feasible alternative, he impedes Victim's return to the pre-coercion setting, slips into into a role not unlike the one Coercer has, and thereby facilitates the ultimate success of the coercion.
Admittedly, there will be controversy over what counts as unreasonable costs,10 why exactly Recipient ought to help Victim and thereby avoid facilitating coercive success,11 and whether Recipient's help should be legally enforceable.12 But at this point, I will not attempt to take a stance on these further questions. I only wish to defend my more general and independent claim that denying Victim a sufficiently costless alternative invalidates consent, and I take Mischievous to illustrate this point. What I can say, however, is that facilitating coercive success could be described, more specifically, as a violation of a duty to rescue, viz. a duty to help someone escape an immediate peril, rather than a general duty of beneficence.13

An Alternative Solution Based on the Modified Coercion Principle
The last section claimed that Recipent invalidates Victim's consent when Recipient (1) exerts or collaborates in exerting the coercion, (2) exploits the coercion, or (3) facilitates coercive success. This section pulls these threads together and presents an alternative solution to the inconsistency of the Coercion Principle and the Permissibility Principle. This solution is based on a distinction between consent being 'obtained by' coercion and consent 'being merely motivated' by coericon. Consent is 'obtained by' coercion, I argue, if and only if Recipient acts as in (1), (2), or (3). 'Obtain' implies the involvement of Recipient and that Recipient uses or illegitimately relies on the constraint on Victim's situation to receive consent to a particular interaction, as is the case in standard two-party scenarios of coercion as well as in my examples Tattoo Money, Tattoo Sex, and Mischievous.14 10 Demanding views could hold that one is obliged to help as long as one does not sacrifice something of (comparable) moral significance (Singer, 1972(Singer, , 2019 or does not jeopardize one's own prospects for a minimally flourishing life (Fabre, 2006, chapters 1 and 2). Less demanding views may only require the acceptance of moderate costs (Barry & Øverland, 2013), only a slight sacrifice (Scanlon, 1998, p. 224), or no genuine costs at all (cf. Ferzan, 2017). 11 For instance, is there a sui generis duty of justice (Miller, 2020) or a duty from an implicit social contract (Valentini, 2015)? 12 See Fabre (2006, chapter 2), Ferzan (2017), and Whelan (1991). 13 I do not talk about a duty of beneficence because beneficence, unlike rescue, is too broad.
As Beauchamp explains, beneficence is understood "to include effectively all norms, dispositions, and actions with the goal of benefiting or promoting the good of other persons" (Beauchamp, 2019). Yet, in cases like Mischievous, the situation is much more specific and closer to situations of rescuing, i.e. helping someone escape an immediate peril. 14 I thank John Hyman for helping me to clarify this distinction.
These cases differ significantly from the original Tattoo case, in which Recipient did not exert coercion, exploit coercion, or facilitate its success. I therefore argue that, in Tattoo, consent was 'merely motivated by' coercion. Consent is merely motivated by coercion if a person's reason for consenting is to prevent a threat from being carried out but the threat came from a third party and Recipient neither contributed to it nor wronged Victim in any of the three ways outlined. Thus, the contrast between 'obtained by' and 'merely motivated by' reflects a shift of perspective: the former looks at how Recipient uses or illegitimately relies on coercion to receive consent and the latter looks at what Victim's reasons for consenting are.
The distinction between consent being obtained by coercion and consent being merely motivated by coercion also disambiguates the Coercion Principle. It is not true that consent is invalid if a person is coerced into consenting. Whether coercion invalidates consent depends on whether consent was, more specifically, 'obtained by' coercion. Adopting this view provides the following alternative solution to the dilemma at hand. It is now the first set of claims that requires change:

1.* Modified Coercion Principle: Consent is invalid if obtained by coercion
and valid if merely motivated by coercion.

2.* Specified Facts in Tattoo:
Victim's consent is merely motivated by coercion.

3.* Modified Conclusion: Victim's consent is valid.
The second set of claims, on the other hand, can be left untouched:

Verdict in Tattoo:
Recipient narrowly permissibly executes the tattoo. 6. Conclusion: Victim's consent is valid.
As a result, there is no longer any contradiction. Modified conclusion (3*) and conclusion (6)  did not obtain consent by means of coercion, notwithstanding the fact that the final agreement between Recipient and Victim was not a reasonable joint decision. After all, Recipient neither exerted coercion nor took advantage of it. Therefore, my view holds that Recipient executes the tattoo narrowly permissibly and reaches the correct result in this case.
Moreover, unlike Millum's solution and any other version of the Influence-Focus-View, my view avoids introducing a concept distinct from valid consent, e.g. a reasonable joint decision or 'request X,' when doing so lacks justification. The Modified Coercion Principle only asks for adjusting the validity condition concerning coercion. In this regard, my solution avoids the second flaw of an Influence-Focus-View and thereby displays greater theoretical parsimony.
For these reasons, my Modified Coercion Principle, as a distinct version of a Recipient-Focus-View, presents a better solution to the inconsistency of the Coercion Principle and the Permissibility Principle than Millum's account. However, the defense of my view is not yet complete: in the next section, I will show that my view presents a distinct contribution compared to other versions of the Recipient-Focus-View before outlining how my view is also able to deal with cases that are insurmountable for other Recipient-Focus-Views. On this basis, I will then finally conclude the support for my view as the correct approach to consent and coercion more generally.

Distinguishing My Account from Other Recipient-Focus-Views
There are the three major alternative versions of the Recipient-Focus-View in the literature. The first comes from David Owens, who states the following about third-party coercion in the context of promises: "where an act invalidates the promise which it induces because it wrongs the promisor, that act must involve the promisee. I leave the notion of 'involvement' here intentionally vague. The promisee is certainly involved if he coerces or misleads the promisor into promising but he might also be involved in some cases where a third party set out to extract this promise by duress or misrepresentation. The crucial point is that the promisee is definitely not involved unless he is the author of the duress/misrepresentation, or had prior knowledge of it. And where he is not involved, the wrong of coercion or misrepresentation cannot invalidate a promise made to him." (Owens, 2007, p. 305, original emphasis) (Owens uses 'duress' and 'coercion' interchangeably.) Owens claims that his position on promises applies equally to consent.15 Thus, Owens accepts a Recipient-Focus-Views as he claims that the validity of consent depends on Recipient's wronging of Victim when obtaining consent, rather than a third party's coercion alone. We also agree on the uncontroversial point that if Recipient exerts the coercion himself, Recipient wrongs Victim and invalidates consent. However, I disagree with Owens on the view that Recipient does not wrong Victim in the relevant sense "unless he is the author of the duress [or coercion] ( … ), or had prior knowledge of it." In fact, my view rejects authorship and knowledge as necessary conditions for invalid consent. By way of illustration, consider the following example: Neglect Victim is a prospective organ donor to her cousin. However, Victim does not want to donate and will only do so because of her uncle's severe coercion. Luckily, Recipient, i.e. the surgeon, could help Victim escape the coercion by issuing a statement on Victim's incompatibility as a donor. According to the hospital's policy, a person will be identified as 'incompatible' due to physiological, psychological, or other reasons. 'Other reasons' are broadly understood and include circumstances of this kind. The exact reasons for incompatibility will be kept confidential within the medical team and not released to anyone except the tested person. Producing such an incompatibility statement is part of general clinical practice and does not break any rules of professional medical conduct. The situation is such that Recipient should have detected the uncle's coercion. Unfortunately, Recipient fails to do so and, as a result, denies Victim the opportunity to speak to him without her uncle being present and thereby also denies Victim the incompatibility statement. Later, Victim consents to donating a kidney.
In Neglect, Recipient is neither the author of the coercion nor did he have prior knowledge of it. Hence, on Owens' account, Victim's consent is valid. Yet, this result is incorrect. Recipient wronged Victim by depriving her of an opportunity to talk to him confidentially and receive a general incompatibility statement. And such wronging invalidates Victim's consent to donation. Unlike Owens' account, my view can explain this result. I explicitly hold that consent is invalid when Recipient denies Victim an alternative to acting on consent (i.e. an alternative to performing the surgery), which would also allow Victim to avoid the threatened harm and not impose an unreasonable cost on Recipient or anyone else. Providing an incompatibility statement is such an alternative, allows Victim to escape the threatened harm and does not impose an unreasonable cost. In fact, given the special fiduciary relationship between patients and physicians, Recipient is obliged to provide such an alternative and help Victim escape the coercion. Otherwise, Recipient unjustifiably facilitates the ultimate success of the uncle's severe coercion. Thus, failing to provide Victim with an incompatibility statement in the circumstances described in Neglect wrongs Victim and invalidates Victim's consent to the surgery. Thus, my account arrives at the correct result and seems therefore preferable to Owens' version of the Recipient-Focus-View.16 The second alternative version of a Recipient-Focus-View comes from Tom Dougherty, who also claims that a person's consent can be valid even if there was coercion from a third party. Dougherty claims: "If X [i.e. Victim] sincerely performs a speech-act that authorizes Y [i.e. Recipient] to perform A as an extension of X's own agency in circumstances that are beyond the control of Y, then Y has X's valid consent to perform A." (dougherty, 2019, p. 11) Applied to Tattoo, Victim's consent is valid on Dougherty's view because the third party's coercion is beyond Victim's and Recipient's control, and Recipient's execution of the tattoo can be seen as an 'extension of Victim's agency.' 'Extension of Victim's agency' means that "If Victim were directly able to control [Recipient's] behavior with their own choices, then Victim would choose that [Recipient execute the tattoo]" (Dougherty, 2019, p. 11). In Dougherty's view, such an 'extension of Victim's agency' is a sufficient (yet not a necessary) condition for valid consent. This is another Recipient-Focus-View as it makes the validity of consent depend on Recipient's action, rather than the third party's coercion alone.
Dougherty's view gives us the right result in Tattoo, i.e. that Victim's consent is valid. Yet, it is too demanding in other cases. Consider the following example: No Hero Recipient has two options to rescue Victim from third-party coercion. Both options involve the execution of a particular tattoo. However, one option comes with great costs for Recipient, i.e. suffering severe physical harm, and the other option comes with no costs for Recipient. If Victim were able to control Recipient's behavior directly with her own choices, Victim would choose that Recipient take the first option as this option involves a smaller and less painful tattoo for Victim. Yet, Recipient cannot be expected to bear such high costs and, as a matter of fact, Recipient only offers the second option. Victim eventually consents to it.
In this case, Recipient's execution of the tattoo is not an extension of Victim's agency in Dougherty's sense, for if Victim were able to control Recipient's behavior directly with her own choices, Recipient would take the first option, not the second. Yet, if Victim consents to the second option and Recipient executes it, Victim's consent is still valid and Recipient acts permissibly. This is because denying Victim the first option does not wrong Victim. There is a limit to the costs that Recipient can be expected to bear for the sake of Victim. Hence, even if Recipient's conduct falls short of Dougherty's 'extension of agency,' consent can still be valid. We should therefore add to Dougherty's view that, not only the extension of Victim's agency, but also Recipient's conduct in No Hero is sufficient for valid consent. My own view can step in here and explain that consent is valid in No Hero because Recipient did not obtain consent by means of coercion, viz. because Recipient did not exert coercion, exploit coercion, or facilitate coercive success as I defined it.
Thus, there are two key differences between Dougherty's view and my view: (i) what I consider to be sufficient for valid consent (i.e. that consent is not 'obtained by coercion') is less demanding than what Dougherty considers to be sufficient (i.e. the 'extension of Victim's agency'), and this is why my view delivers the correct result in No Hero; (ii) unlike Dougherty, I provide both sufficient and necessary conditions for valid consent in cases of third-party coercion. For these reasons, I argue, my view is preferable to, or at least importantly supplements, Dougherty's proposal.
Finally, Mollie Gerver endorses another version of the Recipient-Focus-View, claiming that when "X coerces Y into consenting to Z φ-ing ( … ), Y's consent in such cases is valid if Y is choosing between options that include all those Z has a duty to offer Y, and no autonomy-reducing options Z has a duty to not offer Y" (Gerver, 2021, p. 246, original emphasis). This is a Recipient-Focus-View because it makes the validity of consent depend on how the consent-receiver interacts with the consent-giver, rather than on the third party's coercion alone. Among the various differences between Gerver's insightful account and my view, I will here only concentrate on the most fundamental divergence.
Gerver accepts, and I reject, the view that "where the consent-seeker is aware of the threat, the consent-seeker implies a threat himself" (Gerver, 2021, p. 265). Thus, Gerver thinks that in any third-party case of invalid consent, the consent-seeker issues at least an implicit threat of the form 'if you don't do as I say, I will let you face the harm.' In fact, Gerver makes the following statement about such cases in which Recipient is aware of the coercion: "that agents can fail to obtain valid consent despite engaging in no coercion themselves ( … ) is incorrect" (Gerver, 2021, p. 261).
I disagree with Gerver's requirement of an 'implied threat.' Gerver's view fails because there are cases in which there is no interaction between the consent-giver and the consent-seeker, and, in such cases, there is invalid consent but no implied threat. When Gerver's view is applied to such cases, it is therefore either inapplicable or delivers the wrong result. Since these cases pose a more general problem for Recipient-Focus-Views, I will discuss one central example in the next section and then explain how exactly my view differs from Gerver's view in this regard.

3.4
An Important Further Challenge Consider the following example: Envelope Recipient receives an envelope containing a note with a username and password for online banking, inviting Recipient to log on. The envelope also contains a thumbnail drive. On it, Recipient finds a reliable and accurate video of Victim writing the note at gunpoint. Yet, Recipient's actual logging on is not necessary for Victim to escape being shot. Victim has already escaped Coercer when the letter reaches Recipient, and Recipient knows this. Still, Recipient does not know who Victim is, has no possibility to contact or interact with Victim, and hence no possibility of offering Victim anything.
Here, Recipient cannot permissibly log on and Victim's consent to Recipient logging on is invalid. Envelope is a particularly challenging case for any version of the Recipient-Focus-View precisely because there is no interaction between consent-receiver and consent-giver. It seems that any Recipient-Focus-View, which makes the validity of consent dependent on some wrongdoing of Recipient, needs to identify an instance of Recipient's wrongdoing (or, as Gerver would say, some 'implied threat') that is distinct from and prior to the wrongdoing of acting on consent, precisely because the Recipient's wrongdoing is supposed to explain why consent is invalid and acting on consent is wrong in the first place. Yet, in Envelope, there is no such prior time at which Recipient could wrong Victim. Thus, a successful reply to Envelope will advance our understanding of Recipient-Focus-Views more generally. Unfortunately, scholars, including Gerver, have not addressed such cases. My own view offers a solution. I argued that consent is invalid when Recipient 'facilitates coercive success,' i.e. when Recipient denies Victim an alternative to acting on consent, which would also allow Victim to avoid the threatened harm and not impose an unreasonable cost. Usually, an alternative is another way to escape the threatened harm, and Recipient would normally be able to offer that alternative before Victim consents. However, in my view, an alternative can take other forms too, even the form of an omission. And in Envelope, it does: simply by not logging on can Recipient make it the case that Victim not only avoids the physical injury, i.e. the threatened harm, but also the exposure of the bank account, i.e. what Victim was forced to consent to. Moreover, not logging on does not impose unreasonable costs on Recipient (or anyone else), and Recipient should assume that Victim would ask Recipient not to log on if able to communicate freely. Hence, if Recipient logs on, he 'facilitates coercive success' and consent is therefore invalid.
Using this notion of 'facilitating coercive success,' gives us two different ways of responding to Envelope.

3.4.1
Recipient Invalidates Consent According to a first solution, Recipient commits two wrongs simultaneously when logging on: he facilitates coercive success and acts on invalid consent. These are distinct wrongs because facilitating coercive success, when there is a feasible alternative, is wrongful independently of whether consent is involved.17 Yet, they are not independent wrongs because the facilitation of coercive success explains why consent is invalid in the first place and acting on it wrong. In other words, the facilitation of coercive success may not be temporarily prior to acting on invalid consent, but it is explanatorily prior.
Let me illustrate this point with a story. Greek mythology tells us that the moment King Midas touches food, he turns it into gold and renders it inedible. The food becomes gold and inedible at the same time. Yet, one explains the other: the food is inedible because it is golden, rather than the food is golden because it is inedible. Thus, the story of Midas supports the distinction between temporal and explanatory priority: two things may happen at the same time, yet one explains the other.
One version of my Recipient-Focus-View could propose a very similar picture for Envelope: the moment Recipient acts on consent, he facilitates coercive success and renders consent invalid. Thus, just as the conversion into gold explains the inedibility of food, the facilitation of coercive success explains the invalidity of consent. The fact that everything happens at the same time does not preclude this explanatory order. Hence, this version of the Recipient-Focus-View can deal with cases like Envelope by endorsing the distinction between temporal and explanatory priority, and making only the latter a requirement for wrongdoing to invalidate consent.
Although this view arrives at the correct verdict and remains a version of the Recipient-Focus-View, there is one unconventional implication, namely that Victim's consent is either valid, or indeterminate, up until the moment when Recipient logs on. Those who find this implication implausible can opt for the following alternative, which is also based on my view of 'facilitating coercive sucecss.'

3.4.2
Recipient Re-validates Consent So far, I have assumed that the validity of consent depends on Recipient's conduct in the sense that Recipient's conduct, and not the third party's coercion, invalidates consent. Yet, alternatively, one could also argue that the validity of consent depends on Recipient's conduct in the sense that it can re-validate consent, i.e. make an otherwise invalid token of consent valid again. Applied to Envelope, this view would consider Victim's coerced consent invalid from the beginning and claim that if Recipient acts on that consent without thereby 'facilitating coercive success,' as I defined it, then Recipient re-validates consent, viz. he makes Victim's otherwise invalid consent valid again. In Envelope, Recipient cannot log on without facilitating coercive success, hence cannot re-validate consent. Yet, in another version of Envelope, in which logging on is necessary to save Victim, Recipient may log on without facilitating coercive success and thereby re-validate consent.
This view is a hybrid version of an Influence-Focus-View and a Recipient-Focus-View. On the one hand, it concedes to the Influence-Focus-View that, in cases without interaction beteween consent-receiver and consent-giver, coercion always invalidates consent, irrespective of who exerts it. On the other hand, this view sides with the Recipient-Focus-View in claiming that the validity of consent crucially depends on Recipient, precisely because the validity of consent ultimately depends on whether Recipient facilitates coercive success when acting on consent.18 This hybrid view gives us the right result in Envelope: consent is invalid. Moreover, it remains true to my substantial claims about 'facilitating coercive success' and to the view that the validity of consent crucially depends on Recipient. Finally, unlike the first solution, this hybrid view can describe Victim's consent as invalid from the moment Victim gives it. Yet, this hybrid view also makes a critical concession to the Influence-Focus-View and thereby relinquishes part of Recipient-Focus-Views.
Both of these solutions are based on my account of what it means to 'facilitate coercive success' and consistent with my overall framework. I opt for the first of these two solutions because it is closer to how I defined my view initially and because it does not have to make a concession to the Influence-Focus-View. Yet, readers with a diverging opinion could still adopt the second solution too.
But let me now return to Gerver's view and explain how this discussion of Envelope sets my view further apart from hers. Recall that Gerver maintains that, where consent is invalid and "where the consent-seeker is aware of the threat, the consent-seeker implies a threat himself" (Gerver, 2021, p. 265). By contrast, I deny that an implied threat is needed for consent to be invalid. As a result, Gerver's view and mine differ not only extensionally, viz. they not only give different verdicts on certain cases (e.g. depending on whether there is an implied threat), but they also differ intensionally, viz. they give different explanations for the invalidity of consent in cases of third-party coercion.
Envelope shows why my view is preferable in this regard. In Envelope, consent is invalid but Recipient does not issue any threat, either explicit or implicit. Thus, Gerver's view, unlike mine, is either inapplicable to Envelope or suggests that consent is valid due to the absence of a threat from Recipient.19 The conclusion that consent is valid is unacceptable. But claiming that Gerver's account is not supposed to apply to cases like Envelope is also an undesirable response, for it would render Gerver's view incomplete compared to Influence-Focus-Views, which apply across all these different cases. What is worse, this could suggest that Recipient-Focus-Views might be altogether mistaken if they cannot explain important cases like Envelope. I therefore conclude that my solution to Envelope not only sets my view apart from Gerver's, but also advances our understanding of Recipient-Focus-Views more generally, by demonstrating how Recipient-Focus-Views can deal with especially difficult cases like Envelope.

Conclusion
In this paper, I argued that there is a fundamental conflict between what I called the Permissibility Principle and the Coercion Principle and I proposed a modification of the Coercion Principle as a solution. This solution is based on the view that coercion invalidates consent only if consent was 'obtained by' coercion but not if 'merely motivated by' coercion. I argued that this solution best resolves the inconsistency between the two principles and also challenges how we currently think about consent more generally. I took the current paradigm to be what I called the Influence-Focus-View, according to which there is no relevant difference between two-and three-party cases when identifying those influences that invalidate consent. The validity of consent always depends solely on the features of the exerted influence, but not on who exerts it. However, if my solution based on the Modified Coercion Principle is correct, the Influence-Focus-View is false. Whether an influence invalidates consent does not just depend on whether it is coercive or, more generally, illegitimately controlling. Rather, the validity of consent depends on the person receiving consent. For this reason, I 19 We can also see that Gerver's view has this problem by focusing on the more general formulation of her account. Gerver claims that when "X coerces Y into consenting to Z φ-ing ( … ), Y's consent in such cases is valid if Y is choosing between options that include all those Z has a duty to offer Y, and no autonomy-reducing options Z has a duty to not offer Y" (Gerver, 2021, p. 246, original emphasis). Gerver's focus on what Recipient is under a duty to offer to Victim implicitly assumes that Recipient is always in a position to interact with Victim prior to Victim's act of giving consent. However, this implicit assumption renders Gerver's view unable to deal with Envelope. Since Recipient was not able to offer anything to Victim in Envelope, given Recipient's lack of any possibility to interact with Victim, Recipienta fortiori (on the basis that 'ought' implies 'can') -did not violate a duty to offer something to Victim. Therefore, either Gerver's view does not apply to cases like Envelope or a literal application of Gerver's view leads to the false conclusion that consent is valid.
defended what I called the Recipient-Focus-View, I explained how my particular version of it differs from alternative versions of this approach, and I showed how my view can deal with particularly difficult cases like Envelope.