Consent Under Pressure: The Puzzle of Third Party Coercion

It is a commonplace that coercion by the recipient of consent renders that consent invalid. It is also commonly held that the pressure of natural misfortunes does not invalidate consent that is given as a result. But what about when the coercive force comes from a third party, not from the person to whom consent would be proffered? Consider the following cases, drawn from the experiences of physicians and health researchers while obtaining informed consent.

Teenager. A 16 year old male is participating as a healthy control in an epilepsy study. He is articulate and thoughtful, and clearly understands the study well. When asked why he agreed to participate he replies: “My mom thinks I should do the study while I'm here. If I don't, she'll be mad, so it's easier just to go along with it.”¹

Depressive. Mr. R is blind and suffers from severe unipolar depression that has not responded to medication. His psychiatrist recommends electroconvulsive therapy. During discussion of the pros and cons of treatment, Mr. R confides that he is terrified of ECT, but his family have told him that they will not take care of him at home any longer unless he undergoes it.²

Wife. A research team is recruiting participants for a genetic study of hypertension from patients who attend a health clinic in rural Nigeria. The first woman the researchers ask to participate tells them: “I have to talk to my husband. He makes those decisions.” A day later she returns and says that she is willing to participate.³

In this paper I analyze how threats from a third party affect consent. I argue that, as with other cases of coercion, we should distinguish threats that render consent invalid from threats whose force is too weak to invalidate consent and threats that are legitimate. Illegitimate controlling third party threats render consent invalid just as they do in two party cases. However, knowing that the consent is invalid is not sufficient to tell the recipient of consent what she may or should do. I argue that when presented with a token of consent from someone whom she thinks is experiencing an illegitimate controlling threat, the recipient may act on that token if and only if doing so represents a reasonable joint decision for her and the victim of coercion. The appropriate action for someone faced with third party coercion therefore depends on the other options open to her and those open to the victim of coercion. I close by applying my analysis to a case of consent to medical research.

Coercion and Consent

Recent philosophical analysis of the nature of coercion focuses on interpersonal threats. Though disagreements about some details remain, there is consensus that the paradigm case of coercion involves one person (the coercer) threatening to make another person (the victim) worse off if he does not comply with the coercer's demands.⁴ Thus, when a robber pulls out a gun and says “Your money or your life!” this constitutes coercion because the robber threatens to make her victim worse off (dead) if he does not comply with her demands (handing over his money).

In situations like robbery at gunpoint, successful coercion renders consent invalid. The victim's agreement to transfer his money to the robber does not actually change who has the (legal or moral) property rights over it; that is, the token of consent is not transformative. Similarly, if a surgeon forced her patient at gunpoint to agree to a medical procedure, he would not be giving valid consent to that procedure.

However, not all threats invalidate consent. Suppose the robber were armed only with a sharp tongue and threatened: “Your money or I mock you!” Assuming that is really all that her victim thought she would do, we would expect him to resist such a weak threat. It would be strange for him to claim that he had been forced to hand over his money. Threats that the person threatened could reasonably be expected to resist do not render consent invalid.⁵

Why can powerful threats invalidate consent, whereas weaker threats cannot? The most straightforward explanation has to do with control. When someone makes a powerful threat, like “Your money or your life,” she takes control of her victim's decision. But this is a decision that is the victim's to make, since he is an autonomous person.⁶ Hence, this sort of coercion violates the victim's rights. Control is a matter of degree: we can control our decisions (or other people's) to a greater or lesser extent. In cases where the threat is much weaker, then, the person issuing the threat does not have very much control over her victim's decision. Ruth Faden and Tom Beauchamp say that for consent to be valid the person giving consent must act in a way that is “substantially noncontrolled” (Faden and Beauchamp 1986, 259). This does not rule out some control by other people—some pressure may still allow the person to exercise his autonomy— but it rules out control over a certain level.⁷ In the following, I reserve the term coercion for just those threats that are controlling, that is, sufficiently powerful to render consent invalid. I use the term pressure to include any proposal to make another worse off, whether or not it is controlling.

Sometimes a threat can be controlling, but still not invalidate consent. For example, suppose that Flora is tired of her boyfriend lounging around the apartment all day and not contributing to the rent. “Pay rent or we break up!” she might tell him. If, as a result of her threat, Ivan gets a job and gives Flora rent money, his transfer of money to her is still valid. This is true even though the threat may be controlling: it might be much worse for Ivan that his girlfriend break up with him, than that she break his nose, but the first is permissible and the second a serious violation of his rights. Thus, a legitimate threat will not invalidate consent.

There is an open question about exactly what constitutes a legitimate threat. Some theorists will deny, for example, that it is always permissible for someone to threaten to do something that she has a right to do.⁸ Some, who think coercion always involves a proposal to do something wrongful, will deny that there can be any legitimate threats at all. On such an account, all such apparent threats would really be conditional offers not to do what the person issuing the “threat” had a right to do.⁹ None of these disagreements needs to be resolved in order to analyze third-party coercion; their resolution will just make a difference to the content of the set of threats that are legitimate and so do not render consent invalid, or to the label we give to one set of proposals that have the surface appearance of threats. What I call legitimate threats therefore simply constitute a subset of the proposals that it is permissible for one agent to make to another (the subset of proposals that are experienced as or judged to be threats).

The foregoing discussion implies that there are at least two cases where someone may be pressured by a third party into tokening consent, but where the pressure will not render his consent invalid. First, it will not invalidate consent when the pressure is so weak that the victim's actions remain substantially non-controlled. In this case, even if the third party is doing something wrong, the consent remains valid and ceteris paribus the recipient of the consent may proceed. The Teenager case above probably falls into this category: if there is no reason for the researchers to think that the boy will experience more serious consequences if he disobeys, then they might reasonably expect him to be able to resist his mother's pressure.

Second, third party pressure will not invalidate consent if it is legitimately exercised, even if it controls the victim's decision. This is because the person exercising the pressure has the (moral) authority to control the decision. Again, ceteris paribus the recipient of consent may proceed in the face of such pressure. Depending on the details of the case, the man in Depressive may be in this situation: he is threatened with a consequence that is very serious, but it may be that his family have the right to refuse to take care of him if he will not try the recommended treatments for his illness.

Illegitimate Third Party Coercion

The most interesting cases of third party pressure are those in which someone is threatened with a consequence severe enough to compromise voluntariness and the threat is illegitimate. In a standard two party case, the victim's consent token would then be invalid and the recipient of consent would not be permitted to act on it. I now argue that illegitimate and controlling third party threats also invalidate consent; however, what the recipient of consent may or should do depends on further factors.

The few people who address illegitimate third party coercion do not articulate the justifications underlying their views. Paul S. Appelbaum and colleagues think that consent given under such circumstances is invalid. However, they give no argument for this claim, but seem to simply assume that third party pressure must have the same effect on voluntariness as the same pressure would in a two party case (Appelbaum et al. 2009).¹⁰ On the other hand, Stephen D. Mallary and colleagues assimilate third-party coercion to other strong negative stimuli (such as physical illness) that may lead someone to consent.¹¹ Likewise, Franklin G. Miller and Alan Wertheimer think that the consent in such cases is valid. They appeal to the following case, which I use as the starting point for my analysis:

Pimp. C tells B that he will beat her up unless she earns $500 this evening. B proposes to A that they have sexual relations for $100. A accepts. (Miller and Wertheimer 2010, 96)

In analyzing Pimp they write:

It might be thought that B's consent cannot be transformative if B is coerced into having sex, but that seems much too quick. The standard cases of coercion are those in which the coercer is also the recipient of consent. Here, B is coerced by C, not A, and whatever else we might think of A, we do not think A is guilty of rape. (Miller and Wertheimer 2010, 96)

They claim that this judgment holds even if A knows about C's coercion.¹²

Thus, Miller and Wertheimer seem to think that our intuitive responses to cases like Pimp show that even illegitimate third party pressure cannot invalidate consent. I think their moral verdict on Pimp is mistaken; B's consent is actually invalid. In fact, I now argue, all cases of illegitimate, controlling third party pressure render consent invalid. However, that is not always sufficient to answer the question of what the recipient of consent ought to do.

Note, first, that there are cases of third party coercion where it is clear that the consent is invalid and where that fact makes a difference to the duties of the recipient of consent. Consider the following:

Cash transfer. Coercer tells Victim that he will beat her up unless she gives Recipient $500. Victim gives Recipient the cash, which he puts in his mattress for safekeeping. Later, having escaped Coercer's clutches, Victim asks Recipient for the money back.

Here, the moral verdict seems clear: Recipient ought to give back the money. Moreover, intuitively, the reason he ought to do so is that the cash transfer was not legitimate in the first place. This implies that the consent Victim gave (to the transfer of ownership) was invalid.¹³

There is a simple explanation of why the consent in Cash transfer is invalid. Assume that Coercer's threat is controlling and illegitimate. That means that Coercer is effectively making Victim's decision for her, by leaving her with only one option she can reasonably take, but that he lacks the moral authority to make that decision. His actions cannot therefore constitute an exercise of her autonomy right. The person who has the moral authority—Victim—is not making the decision to exercise that right herself, because she is being controlled. (She decides to do what is reasonable, but that is not the same as exercising a right.) But valid consent requires that the person who has the authority to make a decision make that decision herself. Hence, the consent token is invalid.¹⁴

This explains why the handover of money to Recipient is not legitimate. However, the explanation I gave makes no reference to any facts that are distinctive of Cash transfer: it mentions only features that are common to all cases of illegitimate, controlling third party pressure. It therefore implies that all cases of illegitimate, controlling third party pressure invalidate consent.

Responding to Third Party Coercion

The explanation I just gave, and consideration of cases like Cash transfer might lead us to think that illegitimate third party coercion always makes it impermissible for Recipient to accept Victim's consent token when he knows about the coercion. But that would be too fast. Consider Tattoo:

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.

Here, it seems that Recipient does no wrong by proceeding (and, perhaps, has a duty to do so).¹⁵ It therefore suggests 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. But what explains why Recipient is permitted to proceed? Two factors appear morally relevant: the other options available to Recipient, and what Victim wants Recipient to do. Contrast Tattoo with a case in which Recipient has an option other than proceeding on the token of consent that would be preferred by Victim and that he could easily take.

Pimp without pay. Coercer tells Victim that he will beat her up unless she has sexual relations with Recipient. Victim proposes to Recipient that they have sexual relations in a private room—Coercer will infer what has happened. Recipient agrees.

In this case it seems clear that Recipient ought to go with Victim to the private room, but no more than that. To actually have sex with Victim would be terrible. Indeed, given that he has other options that he could easily take, to have sex with Victim under these circumstances would constitute coercion by Recipient, as well. In effect, he would be threatening to let Coercer carry out the beating if she did not have sex with him.¹⁶

In general, if Recipient has another option available to him, which he could easily take, and which Victim would prefer to him going along with Coercer's demands, he should not proceed on her token of consent. Recipient thereby respects her autonomy by giving her as much control over her consent decision as he is able. If, instead, the option that Victim would prefer is for Recipient to act on her token of consent, as in Tattoo, then he can respect her autonomy by proceeding.¹⁷ In either case, 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.

Consideration of the foregoing cases suggests that there are actually two ways to describe the quality of Victim's consent token in cases of third party coercion. The explanation I have proposed says that Victim's consent is always rendered invalid by third party coercion, but that, depending on the circumstances, what Recipient ought to do with respect to that invalid token of consent may vary. Thus, in Cash transfer Recipient ought to return Victim's money, in Pimp without pay he ought not to have sex with her, and in Tattoo, despite her invalid consent, he ought to tattoo her. An alternative explanation would say that the consent granted to Recipient in Tattoo is valid, since Recipient respects Victim's autonomy as much as he is able. However, her rights are still violated in the same way as if her consent were invalid, since Coercer has made her get a tattoo against her wishes. This interpretation would still say that if Recipient proceeded in Pimp without pay Victim's consent to Recipient would be invalid. Victim's rights would then be violated by two people: Coercer, who creates the threat, and Recipient, who uses the existence of the threat to have his way with Victim.

Both descriptions can diagnose the wrong that is done to Victim (the violation of her rights) and determine what Recipient ought to do. The goals of this paper therefore do not require choosing between them. However, the explanation according to which the consent is always invalid has two advantages that lead me to adopt it here. First, it appears more parsimonious, since it preserves the view that a token of consent is either valid or invalid simpliciter. Second, and more importantly, it captures more clearly the different ways in which the victim of coercion can be wronged and harmed. When Coercer forces Victim to token consent to an act over which she has an autonomy right, he wrongs her. This is true even if the act to which she tokens consent does not actually occur. If that act does occur, then she experiences a harm over and above the wrong of being coerced. For example, if Coercer forces Victim to agree to sexual relations at the point of a knife, then he has wronged her. If he drops dead of a heart attack immediately after issuing the threat then he still did something wrong, but she has been spared the harm of having sex against her will. In the case of third party coercion, describing the token of consent as invalid allows us to capture this distinction. If Coercer forces Victim to agree to sexual relations with someone else, and they have sex, then she will have experienced the harm of having sex against her will, whether or not Recipient is acting permissibly. In the case of Tattoo, Recipient acts rightly by tattooing Victim, but she still has her bodily integrity violated, and she would be better off if Coercer's threat disappeared before the tattooing took place.

Costly Alternatives

We can now see why it is problematic for A to have sex with B in Miller and Wertheimer's Pimp case. If A has $100 to spend on sex with a prostitute, it seems likely that he has $100 that he could do without. Instead of having sex with B for $100, he could just give her the money. Thus, it seems like he has at least one option that he could easily take that B would prefer to having sex with him. Assuming he knows about C's coercion, not to offer her this option would be tantamount to coercing her into sex: just like in Pimp without pay it is as though he proposes to allow her to be beaten if she refuses to have sex with him. Now, this does not entail that A actually has the duty to give her the money, simply because he has been propositioned. It might be that A is also permitted to just walk away from the encounter without giving B anything at all (a question to which I return shortly). However, it does entail that he may not take advantage of her situation to exchange the $100 for sex.¹⁸

Pimp is an easy case, since it is hard to imagine that the recipient of consent cannot spare the $100 that he would otherwise spend on commercial sex. But in other cases the recipient may have alternative options that the victim of coercion would prefer him to take but which are costly to him.

Bike sale. Coercer threatens to beat Victim unless she sells her bike for $100. Recipient is moving to a new city and could do with a bike to get to work; he can't afford to give away $100 because then he will not be able to afford a bike.

Here, we may suppose that Victim would prefer Recipient to just give her $100 but let her keep the bike. But this is not a very good alternative for Recipient: he could give away the $100, but then it would be very difficult for him to get to work. Again, it might be that Recipient is permitted to refuse to transact with Victim altogether. But it also seems plausible that he is permitted to proceed with the sale. How should this be explained?

In cases like Bike sale, the preferences and options of both Victim and Recipient are relevant to the ethical analysis. Each ought to respect the other person, which includes taking account of the costs and benefits for each of their possible decision pairs. Recipient may be permitted to walk away, but he is not permitted to act on Victim's token of consent without proper consideration of what she wants. Victim, on the other hand, may not demand excessive sacrifice from Recipient in order to deal with her coercive situation. If they are going to transact, they therefore ought together to select an option that represents a reasonable joint decision. 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. If they do that, then neither Victim nor Recipient takes advantage of the situation of the other.

Consider how we might work out the most reasonable joint decision in Bike sale. Keeping Coercer's threat fixed and ignoring any other-regarding preferences and values, Victim's options, in preference order from best to worst are:

1. Gift. Victim doesn't give bike to Recipient & receives $100 for Coercer.

2. Sale. Victim gives bike to Recipient & receives $100 for Coercer.

3. Beating. Victim doesn't give bike to Recipient & Coercer's threat is carried out.

Recipient's options, in preference order from best to worst are:

1. Sale. Victim gives bike to Recipient & receives $100 for Coercer.

2. Beating. Victim doesn't give bike to Recipient & Coercer's threat is carried out.

3. Gift. Victim doesn't give bike to Recipient & receives $100 for Coercer.

Victim has no reasonable alternative to Sale or Gift (Beating is ruled out, since the consequences are too bad for her). Recipient has no reasonable alternative to Sale or Beating (Gift is ruled out, since the consequences are too bad for him). Both can blame Coercer for their predicament. But Victim should no more ask Recipient to make the gift than Recipient should recommend that she take a beating. The best option for them to take jointly is Sale, and this is therefore the option that allows both to respect each other's autonomy to the greatest extent possible, given their circumstances. Thus it is permissible for Recipient to proceed with purchasing Victim's bike despite the coercion she is experiencing. The contrast with Pimp here is stark: it is hard to believe that the best joint action in Pimp is for A to give B $100 and for her to have sex with him (that is, if each were taking the point of view of the other as seriously as his or her own, we would not expect this to be the decision they reached). It is therefore implausible in that case that by proceeding on her consent token A respects B's autonomy within the constraints of the situation in which C has placed them.

To sum up, when Recipient knows that he is transacting with a victim of third party coercion, he becomes able to respect or disrespect Victim's autonomy in a way that he could not before, because he understands the constraints under which she is operating. He can act with her by deciding jointly what they should do. If a reasonable joint decision involves him acting on her token of consent then it is permissible for him to do so. Alternatively, he can usurp her decision just as Coercer has by making use of the threat she is under to obtain something to which he has no right. Either way she will be coerced by Coercer. But Recipient's actions determine whether or not he also coerces her.

Note that the analysis of how someone may or should act in the face of consent she knows to be invalid may be complicated by other moral considerations in addition to the rights and interests of the parties to the transaction. For example, some people argue that some deceptive research studies—in which valid consent is not obtained because the research participant does not know what she is agreeing to—can be justified by the social value of the data that is collected (American Psychological Association 2010, 8.07). Here, the gain to society of the knowledge generated is thought to outweigh the wrong of proceeding on an invalid token of consent. Such considerations should already be included in the analysis of the permissibility of proceeding on a token of coerced consent. In a two party case, if it is permissible to proceed on a coerced token of consent, then that coercion will normally be legitimate.¹⁹ In the case of third party coercion, the coercion might be unjustified and yet there might also be moral considerations that favor accepting the consent. For example, significant benefits to others might render it permissible for the recipient to proceed, even though he knows that the victim has been coerced. If such considerations apply to a case, they should also be taken into account when working out what would be a reasonable joint decision: reasonable here does not mean only prudential, but includes each party's moral reasons, too.²⁰

Walking Away

The analysis so far has focused on the conditions under which it is permissible to accept a token of consent from someone who is experiencing third party pressure. Sometimes the potential recipient has low cost alternatives to acting on the victim's token of consent, which she would prefer him to take, and he is therefore not permitted to accept her token of consent. However, I alluded above to the possibility that he might still be permitted to walk away from the interaction altogether, even if that is not the option she would prefer. Under what conditions is this permissible and under what conditions does he have a duty to take one of the other options available to him?

The answer to this question depends upon the other duties that the situation imposes on the potential recipient of consent. For example, the emergency in which the woman in Pimp without pay finds herself and the low cost to Recipient in helping her fool Coercer might make us think that Recipient ought to help her out on the grounds that this would constitute an easy rescue. In a case in which researchers are soliciting people to enroll in their study, they might have some responsibility for the instances of coercion that result, and so have a duty to help their potential participants avoid negative consequences. Doctors and researchers might also have role-based duties or duties arising out of existing relationships with patients that entail that they are not always permitted simply to refuse to engage with the coercive situation. In other cases, there may be no duties that bind the potential recipient of consent or he may have countervailing reasons not to get involved. For example, the tattoo artist in Tattoo might argue that he would be compromising his integrity by tattooing someone under protest, and so refuse to be a party to the coercion.²¹ A complete analysis of the various reasons that might apply is beyond the scope of this paper.

Two Objections

It might be objected that the picture of shared decision-making that I have sketched is unrealistic. In many cases, even if the recipient of consent knows about the third party coercion, she may not have the opportunity for an extended discussion with its victim. She may therefore have limited knowledge of the victim's preferences. Moreover, even if the two have the opportunity to decide together what to do, there is no guarantee that the decision-making procedure will proceed in the way I have suggested. For example, Victim in Bike sale might take advantage of Recipient's willingness to compromise and manipulate him into excessive sacrifice by playing up the awfulness of her situation.

Certainly, these are sensible practical concerns. However, they do not represent a flaw in the ethical analysis. In the absence of an actual discussion, we can still ask what each person would do, if acting in good faith (that is, trying not to take advantage of the other). Recipient should then proceed according to how he judges such a discussion would conclude and can thereby do what he can to respect Victim's autonomy without martyring himself. Cases where communication is difficult or impossible, or joint decision-making is otherwise challenging, will have practical implications for how far each person should go in accommodating the other, but are not problems with the underlying theory.²²

Second, someone might object that my conclusions are unfair to potential recipients of consent. If I am right, someone who discovers third party coercion may, through no fault of her own, have duties to act or refrain from acting in ways that she would otherwise not.²³ For example, suppose we are persuaded that there is nothing intrinsically wrong with exchanging sex for money. Then, when he discovers the coercive nature of the situation, the potential client in Pimp moves from being permitted to transact freely with the prostitute, to being forbidden to do so. In Pimp without pay, just because of the accident of his interaction with Victim, Recipient seems to have acquired the positive duty to go along with Victim so that Coercer will infer that they had sex. Since Recipient did nothing wrong, is it not unfair to single him out for these additional duties?

In response, note that the world is frequently unfair in just this way in other contexts. The innocent passerby who witnesses a crime can thereby acquire a duty to report what he saw in court. Whoever happens to walk by the shallow pond where the child is drowning thereby acquires a duty to rescue him, but, again, through no fault of hers (Singer 1972). There are also other situations in which someone's ability to give valid consent is compromised through no fault of the recipient of consent, but where knowledge of the compromise changes the recipient's duties. For example, a surgeon may have innocently believed that the patient being referred to her was capable of making his own decisions. When she realizes that he is cognitively compromised and does not understand what is involved in the operation she was intending to perform, she is no longer permitted to go ahead with it. Moreover, it is plausible that her discovery entails further positive obligations, for example, that she ought to locate a suitable surrogate decision-maker for him, attempt to explain his condition and the possible treatments available at a level he can understand, and so forth. All of this is a product of her coming to know that someone she had originally had good reason to think had given valid consent, in fact failed to do so.

Consent to Health Research Under Pressure

Cases of consent to health research participation (or the specific procedures that are involved) illuminate some of the complex ethical problems presented by real situations of third party coercion. The primary goal of research, in contrast to clinical care, is the generation of socially valuable data, not the well-being of the research participants. Thus, research participation often involves risks to participants that are not balanced by therapeutic benefits. However, health research may provide information of great value to the populations being studied, and enrollment does sometimes offer a benefit to individual participants that they could not obtain outside of the trial. There may therefore be good reasons for well-intentioned researchers not to simply take the precautionary measure of excluding from research all members of populations at risk of third party coercion, but to consider whether and when their consent can be accepted.

Wife nicely illustrates the various considerations that are relevant to a decision about accepting consent under pressure. First, it is certainly possible that the Nigerian woman's husband backs up his authority with threats that are both illegitimate and controlling enough to undermine the validity of his wife's consent. However, the testimony quoted in Wife does not tell us that for sure: it might be that asking permission is just a part of the shared decision-making that characterizes her marriage. In general, if someone is innocently ignorant that the consent she is offered is invalid, she does no wrong in acting as if it were valid.²⁴ But in cases like Wife there may be reason to suspect that something is going on, but the potential recipient of consent does not know the whole story.

Hence, second, the researchers have a duty to find out more. They need to have some understanding of the woman's culture, since understanding the way that the communities and families in this population are structured, and the way that husbands and wives are expected to interact, may help them understand what it means for the wife to ask her husband about whether to enroll in research and for the husband to say that she should. However, no matter what the cultural mores, every relationship is different. The researchers also need to find out something about how this woman, in particular, relates to her husband and how they make decisions. This may allow them to ascertain whether her husband issues illegitimate threats of sufficient force to invalidate her consent.²⁵

The difficult epistemic situation of the researchers is distinctive of third party coercion. When the coercer is also the recipient of consent, she normally knows about the coercion. But this is not true when third party coercion is involved. Each of the three cases described in the introduction could have occurred without the researcher or physician suspecting any pressure. Moreover, when the perpetrators of coercion expect others to disapprove, they are likely to try to conceal the coercion, which exacerbates the difficulty for a recipient trying to work out what to do. Worse, investigating whether coercion is taking place might also risk putting the victims of coercion in greater danger; for example, if they might be punished by the coercer for arousing the suspicion of the recipients of consent.

Third, what the woman wants to do regarding research participation independent of any threats she is under is crucial. If she originally wanted to participate, and simply got approval from her husband to do so, then she might be living in a state of deplorable subjugation, but her research enrollment would not be involuntary.²⁶ On the other hand, if she did not want to be in the research, but is coming to enroll just because her husband told her to, then the researchers may have a duty to explore alternatives to enrolling her.

Finally, then, if she is the victim of third party coercion the researchers must consider their duties to her, the options open to them, and discuss with her how she would like to proceed in the light of all these options. Since their request for blood samples in the health clinic led to this particular instance of coercion, it is plausible that the researchers have at least some obligation to help the woman avoid the negative consequences of disobeying her husband. If it is possible to exclude her from participation in a way that will not lead to punishment, then the researchers may have a duty to do that. For example, they could say that she was unable to participate because she did not meet the formal inclusion criteria of the study (for example, her hypertension was not severe enough). Alternatively, if the husband is insistent on her participation, then they might have a duty to fake her participation or to allow her to do so. They might give her the Band-Aid that would cover the mark she would have if she had given blood, provide her with the appropriate token of compensation for participation, and so forth.²⁷ Only if it is not possible to find such an alternative that she would prefer, and which would shield her from her husband's punishment, should the researchers enroll someone into their study who is being coerced to participate.²⁸

Conclusion

Third party pressure makes the normative situation of the recipient of consent different from the situation she would face if she herself was the source of the pressure and different from the situation she would face if it were just natural misfortune that put the victim into a bad choice situation. If the pressure is both illegitimate and controlling then it will invalidate consent, just as it would in a two party case. However, unlike in the case where the recipient of consent is also the coercer, sometimes the recipient is permitted to proceed upon this invalid token of consent. She may proceed if doing so would constitute a reasonable joint decision made by her and the victim of coercion. On the other hand, were it just natural misfortune that put the person proffering consent into a bad choice situation, then his consent would be valid, and the recipient of consent would always be permitted to proceed. In the case of third-party coercion, however, if there are options open to the recipient of consent that she could easily take and which the victim would prefer, then she may not proceed, though she might be permitted to walk away.²⁹