Anda di halaman 1dari 30


Manipulation and Wrongdoing Ryan W. Davis Draft: March 2012 Ordinarily, we wrong others when we coerce or deceive them. So prominent are these two personal moral trespasses that it is tempting to think, suitably understood, that they can account for the whole of wrongdoing. Christine Korsgaard has described lying and coercing as the two fundamentally wrong things you can do to others.1 They are fundamental, on Korsgaards view, because they can together explain more complex forms of treating others badly. The failure to meet a friend for lunch after agreeing on a place and time is a species of deception.2 A parents refusal to refrain from interfering in her adult daughters life is wrong because of its coerciveness.3 And so on. Coercion and deception, in turn, can be explained as instances of a single underlying phenomenon: using another as a means.4 Though discerning exactly what using someone as a means involves has long been the subject of some puzzlement, the Kantian explanation of wrongness has clear appeal. By gathering all instances of wrongdoing as violations of what Korsgaard calls the unconditional value of free rational choice, Kantianism offers a unified account of what we owe to others. Notwithstanding its virtues, it is unclear that the Kantian explanation can answer the problem it sets out to solve. Perhaps to their (moral) credit, I think that the inability to Kantian

Christine Korsgaard, Two Arguments Against Lying, in her Creating the Kingdom of Ends (Cambridge: Cambridge University Press, 1996): 335-362, at p. 347. 2 For discussions of deception, see Alan Strudler, Deception Unraveled, Journal of Philosophy 102:9 (2005): 458-473; J.E. Mahon, Kant on Lies, Candor, and Reticence, Kantian Review 7 (2003): 101-133, and Kant and the Perfect Duty to Others not to Lie, British Journal for the History of Philosophy 14 (2006): 653-685; Paul Faulkner, What is Wrong with Lying? Philosophy and Phenomenological Research 75 (2007): 524-527. 3 Coercion has been extensively discussed in moral and political philosophy. For a few classic and recent examples, see Korsgaard, Kants Formula of Humanity, in Creating the Kingdom of Ends, 106-132; Robert Nozick, Coercion, in Philosophy, Science, and Method: Essays in Honor of Ernest Nagel [ed.] Sidney Morgenbesser, Patrick Suppes, and Morton White (New York: St. Martin's Press): 440-472; Alan Wertheimer, Coercion (Princeton: Princeton University Press, 1987); Arthur Ripstein, Authority and Coercion, Philosophy & Public Affairs 32:1 (2004): 2-35; Arash Abizadeh, Democratic Theory and Border Coercion: No Right to Unilaterally Control Your Own Borders, Political Theory 36:1 (2008): 37-65. 4 Korsgaard, Two Arguments, p. 347.

philosophers like Korsgaard to think of more ways to do wrong betrays a certain lack of creativity. It does not take much deviousness to imagine what it would be like to be wronged or to do wrong without coercion or deception. I will argue, at least, that such examples are commonplace. A woman stranded in a pointless meeting is apt to resent its organizer for making her attend, even if no coercion was threatened, no penalties were attached to nonattendance, and she had been put under no illusions about the meetings usefulness. A couple whose able-bodied children were repeatedly called upon to assist their hapless neighbors with yard work might feel indignant about their family being usedeven if no coercion or deception was even lightly applied. In times like these, it is easy to wonder whether our annoyance is not better directed against ourselves for going along with the plans of others. Worse, our resentment might be fanned by our inability to detect in our associates anything warranting resentment. Some people seem so good at being bad that they cannot be blamed for anything at all. Does this mean they do nothing wrong? It would be nice to think that we have something against people who shrewdly get us to do their bidding without the unsubtle tools of coercion and deception. Often a way station on the road to conceding that no wrong has been done is to accuse such people of being manipulative. However, it is difficult to make this accusation stick. A general problem confronted by accounts of manipulation is to explain its wrongness in a way that keeps the reasons for its impermissibility from becoming parasitic on prohibitions against deception and coercion. Usually, manipulation is said to be wrong either because it is deception-like, or because it is coercion-like. On the one hand, manipulators are thought to do wrong by confusing, mesmerizing, or otherwise preventing their victims from grasping the facts about what it would be best to do. On the other, manipulators are sometimes criticized for issuing threats that dont quite rise to the level of coercion (If you dont like my paper, Ill be crushed!). In either case, the manipulee is made to do something that it does not make sense for her to do, either because she is confused about her options, or some of her options are

burdened. I believe that there is a form of manipulation about which moral caution is needed, but that involves no quasi-deceptive or quasi-coercive strains. The reason for this is that it sometimes makes the most sense for the manipulated person to do just what the manipulator asks. One need not be confused or pressured to be manipulated. Manipulators can give us real normative reasons to act when they engage in manipulation, but these are just reasons that they ought notor at least, not usuallyissue to others. Manipulation is wrong not because it induces the manipulee to do something that she should not, rationally, do, but because it takes unwarranted advantage of a moral relationship.

1 In the television program Gilmore Girls, Lorelai, a single mother, attempts to raise her daughter Rory while keeping her own intervention-prone parents at arms length. On one occasion, Lorelais mother, Emily, decides that it would be better if Rory developed a closer relationship with her grandfather. Accordingly, Emily proposes to both her husband and her granddaughter that they spend some quality time with each other on the golf course. When Lorelai learns of her mothers plan, she reacts with anger. In the lines that follow, she tries to say what, precisely, about her mothers intentions she finds so nettlesome. LORELAI: Don't do this, Mom. EMILY: Do what? LORELAI: Force Rory and Dad to go golfing. EMILY: I'm not forcing anybody. LORELAI: Well you're manipulating the situation in a way that gives no one a way out. That's force. Look it up. EMILY: I'm just trying to help your daughter get an education. LORELAI: Thank you. She'll find another sport. EMILY: Why should she? LORELAI: Because she doesn't want to go and Dad doesn't want to take her. EMILY: Oh, your father doesn't know what he wants. He'd get his hair cut at the butchers if I let him. LORELAI: Let it go please. EMILY: Well why don't you just let Rory decide? LORELAI: Because Rory is the sweetest kid in the whole world and she won't tell you that she doesn't want to go because she's too afraid of hurting your feelings.

Lorelais complaint is that Emily is forcing Rory and her grandfather to go golfing together. But as Emily points out, she hasnt used anything that we would ordinarily think of as force.5 She has simply asked her husband to invite his granddaughter to the golf course, and asked Rory if she would like to go. No threatimplicit or announcedis issued against either. Nor has Emily done anything to artificially inflate the desirability of golf in the eyes of her granddaughter. If anything, it seems she has only given her family members a choice that they are free to consider in a clear-headed way. And yet, Lorelai does not interpret Emilys invitation as a kind turn. Her family is caught in a position where they must do Emilys bidding, thanks to her manipulating the situation in a way that gives no one a way out. But one might wonder, along with Emily, how a person could have no way out of a situation in which she understood all of her options and was not coerced into taking any of them. Of course, a family member suspicious over Emilys history of getting her way might fear that there is something coercive up her sleeve. Perhaps she will be stingier in her support of Rorys education or less gracious with her husbands business clients if her offer is rejected. If so, Lorelais complaint could be understood as a combination of accusing coerciveness (the future enforcement of penalties) and deception (not being forthright about what penalties were in place). However, this is not Lorelais reply. She gives no indication that she fears retribution or suspects deception. The offer is exactly what it appears to be: an invitation posed with no attached costs. Lorelai nevertheless believes that Emily is manipulating the situation to get what she wants. To her, Emilys wrongful interference is obvious; one need only look up the definition of force to see that this counts as an example.

She has merely given Rory and Richard an additional optionthe option to golf togetherwhich they are both free to decline. For this view on coercion and force, see Ian Carter, A Measure of Freedom (Oxford: Oxford University Press, 1999).

It is not so clear, however, what dictionary Lorelai has in mind.6 On a fairly standard account, Emily isnt forcing anyone. Philip Pettit, for example, argues that to interfere with a choiceis always to put an obstacle in its way intentionally, or at least in such a manner (say, such a negligent manner) that blame may be in order. This is not to imply that interfering with a choice entails removing it. As Pettit explains:
Interference may certainly involve removing an option from a set of otherwise available options (say, reducing options A, B, and C to options A and B), thereby rendering the choice of that option impossible. But, plausibly, it may also involve changing the options by adding a penalty to one of the alternatives; this might reduce the set to A, B, and C-minus, where C-minus refers to C with a penalty. Equally plausibly, it may mean misleading the agent about the options 7 available.

Here again, coercion and deception are noted as ways of interfering, and Emily is guilty of neither.8 To see why, notice that Rorys option set after Emilys offer is exactly as it was before, with one addition. No options have been removed, and no penalties have been attached to existing options. If Emily is interfering with Rorys choices, it must be because there are more than two ways to interfere with a person.

2 Looking ahead, my effort will be to try to understand how it is possible to wrong someone without coercion or deception. Before turning to that problem, it will help to better understand how coercion and deception are wrong. In this section, I will treat both as violations of autonomy. Autonomy admits of many interpretations. My use will aim at being ecumenical, but one distinction among conceptions of autonomy will be important. This is the difference between

Felicia Ackerman, for example, considers four dictionary definitions of manipulative in her investigation of the concept, finding that none satisfactorily captures the spectrum of cases in which manipulation seems, intuitively, to be wrong. See Ackerman, The Concept of Manipulativeness, Philosophical Perspectives 9:AI (1995): 335-340. 7 Philip Pettit, Freedom in the Market, Politics, Philosophy & Economics 5 (2006): 131-149, p. 135. 8 Throughout, I refer to actions as coercive when they intentionally burden or remove anothers options. Paradigm cases of coercion involve actual or threatened physical force or violence, as I will note later.

what might be called deontic and telic understandings of autonomy.9 These two conceptions refer to two distinctive ways in which autonomy can be valuable. According to the evaluative interpretation, autonomy is valuable as an important component of personal well-being. A life that is autonomous is, all else equal, more likely to be a good life. If I can exercise my capacity to make choices according to my own values (moral autonomy) or according to what I believe there is most reason to do (rational autonomy) or act free from the interference or domination of other agents (political autonomy), then my life will likely go better from my own point of view than it would if I lacked autonomy.10 I will not say anything here about which of these domains is most valuable. The point is just that the role played by autonomy is that of making ones life go well. Contrast this with what I will call deontic autonomy. According to this view, autonomy should be understood as a standing or authority to issue claims to others. Deontic autonomy holds that if I am an autonomous moral agent, I can demand to be treated by other such agents according to certain standards. If I am not so treated, I am warranted in resenting and holding accountable those responsible for my mistreatment. Like evaluative autonomy, deontic autonomy is a value, but the response made appropriate by the value differs. Consider my right to refuse medical treatment if I so choose. One explanation for my right is that if I am permitted to live a life in accordance with my own values when making personal medical choices, then my well-being would be better than a life in which I was denied this freedom. A different explanation would assert that I had a standing to make my own medical choices, and that denying me freedom of choice in this area would fail to acknowledge my authority as a source of claims. In the first, the content of the reason to grant me the right is ultimately given by facts about my well-being. In the second, the content of the reason is ultimately given by the

Cf. Stephen Darwall, The Value of Autonomy and Autonomy of the Will, Ethics 116 (2006): 263-284. See also Philip Pettit and Michael Smith, Freedom in Belief and Desire, Journal of Philosophy 93:9 (1996): 429-449. 10 See ibid., p. 265. For a helpful discussion of different types of autonomy, see Nomy Arpaly, Unprincipled Virtue (Oxford: Oxford University Press, 2003): Ch. 4.

conjunction of (1) my demand, or my disposition to demand, that I be granted the claim; and (2) that I have a standing or authority to issue such a demand.11 Evaluative autonomy explains the right in terms of welfare-based facts. Deontic autonomy reverses the order of priority: my welfare is improved when my warranted claims to direct my own life are not infringed.12 In this paper, I will have in mind the deontic version of the value of autonomy. I will not try to argue that it is superior to the evaluative account, or that the evaluative version is irrelevant to moral reasoning. I will only note that I dont believe the evaluative view of autonomy can tell the whole story. If it did, then our reasons for honoring autonomy would ultimately be reasons of care, or concern for well-being. But this often does not seem to be what we have in mind. Autonomy is more naturally connected with reasons of respect, or recognition of authority that limits or constrains our treatment of another. This is why paternalistic treatment of a moral agent can still be wrong even when it ultimately promotes her welfare.13 At least, I will take this for granted for now.14 If the value of autonomy is to be understood as deontic, a further question arises: What do autonomous persons have authority to demand? As the medical rights example suggested, one frequently suggested candidate is that persons have the authority to demand to determine their own actions. Overriding or interfering with anothers capacity to decide how she will act fails to respect her autonomy. Call this the authorship principle:


This may not be exactly what Darwall considers the content of a second-personal reason to be. For a discussion of this point, see R. Jay Wallace, Reasons, Relations, and Commands: Reflections on Darwall, and Darwall, Reply to Korsgaard, Wallace, and Watson, both in Ethics 118 (2007): 24-36, 5269. 12 For an informative discussion of well-being, see Scanlon, What We Owe to Each Other (Cambridge: Harvard University Press, 1998): Ch. 3. 13 Cf. Darwall, The Value of Autonomy, p. 280. 14 Note that accepting deontic autonomy does not entail a commitment to a non-consequentialist theory of normative ethics. One might believe, for example, that the reasons for respecting the autonomy of others as a deontic standing are underwritten by a practice that is justified by its overall consequences. In other words, accepting deontic autonomy does not rule out consequentialism as criterion of rightness, but it does rule out consequentialism with a decision procedure. I think this is acceptable to many consequentialists.

In the absence of some special justification, an action is morally impermissible if it undermines another persons authorial relationship to her actions.

The normative force of treating persons as the authors of their own actions is underwritten by the importance of autonomy, or in an older, more evocative language, by the dignity of persons as ends in themselves.15 I will not try to argue for the authorship principle here, but I will try to clarify what it means.16 To a first approximation: an agent (A) is the author of her action () to the extent that, were she to know all of the non-moral facts about her -ing, she would not reasonably judge that any other agent shared responsibility with her for the decision to . To respect someone as the author of her actions is to refrain from interfering in a way so as to usurp responsibility for the persons decision to act without her consent. I will try to clarify over the remainder of the section. This idea is a fairly familiar one. Barbara Herman suggests that something like the authorship principle could generally explain the wrongness of interference with persons. She writes:
A law of rational agency that entails the causal control of one will over another implies that no will is a possible source of reasons all the way down. Such a law could not be a law of rational 17 agency: a law describing the agency of ends-in-themselves.

Coercion (ordinarily) violates the authorship principle because it represents an overt attempt by one will to gain causal control over another. Without some special justification to that person legitimating the coercion, coercion is thereby rendered impermissible. Deception similarly disrespects a persons agency. At least one form of deception succeeds by causing a person to perform an action that is different from the action that is intended. For example, a parent might

For recent examples see Adrienne Martin, How to Argue about the Value of Humanity, Pacific Philosophical Quarterly 87:1 (2006): 96-125; William Nelson, Kants Formula of Humanity, Mind 117:465 (2008): 85-106. 16 The idea of understanding action through the metaphor of authorship is suggested in Tamar Schapiro, Three Conceptions of Action in Moral Theory, Nous 35:1 (2001): 93-117. 17 Barbara Herman, Leaving Deontology Behind, The Practice of Moral Judgment (Cambridge: Harvard University Press, 1993): 113-131, p. 126.

deceive her child into participating in a family event by lying about what it would involve. If I deceive you, I displace you as the author of your action.18 Deception denies you the chance to consent to the action that you perform.19 Like coercing, deceiving works to subvert the agency of another, but it goes about achieving that subversion by altering how options are visible, rather than by altering the options themselves. It must be acknowledged that talk of subverting anothers agency, or exercising causal control over anothers actions can be frustratingly opaque. Why, one might wonder, does coercion or deception keep a person from authoring her actions? If I attach a penalty to action C, I still leave you as the person who will choose whether you take option A, option B, or option C. Likewise, if I deceive you about what options are available to you, you will remain the person making the choice about what you will do and not do. You will be making that choice in admittedly worse epistemic circumstances, but this does not imply that it is not you who is acting. Moreover, it is possible that the obstacles to successful action imposed by either coercion or deception could also be created by non-human factors. In deciding what to do, I might have beliefs equally as mistaken as those of someone deceived, but my false beliefs might be a result of natural uncertainty or complexity (think of a road sign that had fallen down). Likewise, I might find that some of my options have been burdened or removed (think of a tree that had fallen across the road I was traveling). If these natural problems do not jeopardize my standing as the author of my action, why should coercion or deception? 20 Here it is important to distinguish between two possible concerns I might have about any action I perform: my expected success in the actions achieving its ends, and my control over determining what the action is. The authorship principle is not meant to imply that a person is more an author of her actions insofar as she can reliably expect them to be successful at
18 19

Cf. Rae Langton, Duty and Desolation, Philosophy 67 (1992): 481-505. Christine Korsgaard, The Right to Lie: Kant on Dealing with Evil, in Creating the Kingdom of Ends (Cambridge: Cambridge University Press, 1996). 20 Sarah Buss emphasizes this objection in her Valuing Autonomy and Respecting Persons: Manipulation, Seduction, and the Basis of Moral Constraints, Ethics 115 (2005): 195-235, pp. 213-214.

achieving her ends. It is not the fact that a decision is less good than it might be that is troubling, but the fact that a decision is less mine than it might be. These concerns are actually quite different, and the difference cannot be discerned from facts about my personal success in acting, nor in facts about my welfare. 21 Rather, what is special about coercion or deception has to do with my relationships with those who coerce or deceive me. In interfering with my choice, the coercer or deceiver seizes responsibility, to some degree, for my action. That is, she fails to treat me with respect. When treated this way by others, we can appropriately resent them for their denial of our authority as persons to make choices for ourselves. To coerce or to deceive is to fail to recognize another as a source of reasons. Because they follow from one persons addressing a claim directly to another, Stephen Darwall refers to such reasons as second-personal.22 As all persons are owed respect, anyone is entitled to resent unjustified coercion or deception. Including second-personal reasons allows a better understanding of how being coerced or deceived differs from being subject to natural obstacles. Imagine two peopleTom and Wendywho are both trying to get to an ice cream shop in the country. Neither of them succeeds: Tom, because he is a poor navigator and misreads his map of the area; Wendy, because she is deceived about the location of the ice cream shop by her paternalistic roommate, who thinks touring ice cream shops is a bad practice for her. From a third-personal point of view, Wendys and Toms plans go awry in the same way. Second-personally, however, Wendy has a claim that Tom does not. If Tom were to know all of the facts about his action (the information on which he formed his intention, and so forth), he would see that no other agents determined what he would do. But if Wendy knew all of the facts about her action, she would see that another agent was partially (in this case, crucially) responsible for what she


For a helpful discussion distinguishing between imposed and natural obstacles to freedom, see Philip Pettit, Republican Liberty: Three Axioms, Four Theorums, in C. Laborde and J. Maynor [eds.] Republicanism and Political Theory (Blackwell, 2008). 22 Darwall, The Second Person Standpoint.

did. While he has only himself to blame, she canliterallyblame someone else. Coercion and deception may not be generally worse for agents than other kinds of obstacles to acting as they choose, but this is not what makes them wrong. They are wrong because they assert control as partial author of someone elses actions.

3 Lorelai accuses Emily of manipulation, which I have understood to mean that Emily committed some wrong that was neither coercive nor deceptive. Her accusation thereby diverges from accounts accepted among philosophers, who tend to see manipulation as either a kind of sophisticated coercion, or else as a variety of sophisticated deception. An offer is manipulative, we are told, when it doesnt let you think about whether to accept or reject it.23 It may be made in the passion of the moment, or the manipulator may know it to be seductive, alluring, or beguiling.24 A department colleague may manipulate you by asking for help when you are distracted and unable to pay full attention to the request.25 Tom Sawyer manipulates his friends when he feigns enthusiasm at the prospect of whitewashing a fence so as to induce them to do it for him.26 Ayn Rand manipulated her protg and his wife into allowing an affair with her by lavishing praise on their individualist ethos.27 A parent might manipulate an adult child by threatening suicide, should the child move away.28 While it might make sense to refer to all of these cases as instances of manipulation, I believe they miss out on the feature of manipulation that makes it morally distinctive. This is because the wrongness of the manipulative action in each of these cases is parasitic on an
23 24

Sarah Conly, Seduction, Rape, and Coercion, Ethics 115 (2004): 96-121. Eric Cave, Whats Wrong with Motive Manipulation? Ethical Theory and Moral Practice 10 (2007): 129-144. 25 Patricia Greenspan, The Problem with Manipulation, American Philosophical Quarterly 40:2 (2003):155-264. 26 Ibid., 155. 27 Marcia Baron, Manipulativeness, Proceedings and Addresses of the American Philosophical Association 77:2 (2003): 37-54, p. 39. 28 Thomas Hill, The Importance of Autonomy, in Autonomy and Self-Respect (Cambridge: Cambridge University Press, 1991): 43-51, p. 48.

explanation of the wrongness of coercion or deception. To show this point, I will suggest that such cases can be carved quite neatly into instances of coercive manipulation, and instances of deceptive manipulation. Coercive manipulation involves the attaching of penalties to some of the manipulees options. It differs from outright coercion either in that the penalties attached are in some way less serious, or in that they are intended but not announced. Consider the case of the elderly parent who threatens suicide if he is not attended to by his adult children. The parent is certainly not threatening his children with violence or physical force, so he is not coercing them in the sense of coercion understood in paradigm cases, like political coercion. However, he is attaching a serious penalty to noncompliance with his requestnamely, the denial of a valuable relationship between parent and child. Sometimes actions labeled as manipulative display explicitly coercive elements. Patricia Greenspan describes practices of bullying or mobbing in the workplace as manipulative.29 However, she also notes that these practices might just as well be described as coercion outright, and suggests what she considers to be a more purely manipulative case, in which employees at a company are made to speak as though they had volunteered for tasks that they actually understood to be compulsory.30 Even this case does not look like pure manipulation though. Requiring another person to play along as though an unfavorable policy treated her well would likely feel especially demeaning, but this requirement could only be enforced through the threat of punishment. The policys perceived coerciveness is crucial to its success. The same conclusion can be reached in considering deceptive manipulation. By deceptive manipulation, I mean to refer to efforts to bring it about that a victim has false beliefs about what she should do, where the efforts to bring about these beliefs do not count as fullblown deception. Tom Sawyers effort to convince his peers of the joy that comes from

29 30

Greenspan, pp. 162-163. Ibid., 163.

whitewashing a fence is a good example. So is the case of the department colleague who asks for help when you are distracted. In the first, Tom believes that whitewashing a fence is not fun, but wants his friends to come to believe (falsely) that it is. In pretending to be excited about the prospect of whitewashing a fence, Tom might refrain from telling any lies. But Tom also does not believe that his friends will actually find whitewashing a fence to be rewarding. He is not sincerely recommending to them an activity that he believes that they would enjoy, even though he happens not to like it. Rather, he takes himself to be inducing them to do something that they have most reason not to do. Tom deceives his friends by reticence, if not by lying. 31 Something similar might be said of the manipulative department colleague. Unlike Tom Sawyer, he does not engage in any antics designed to disrupt your rational consideration of what it would be best for you to do. However, he does take advantage of vulnerabilities in your deliberative outlook. Moments when you are distracted present occasions when you might have let down your guard about what you should do, and thereby be inclined to accept a task that, given more time to consider, you would not have gone along with. Catching you when you are off guard is just one way of bypassing your rationality in securing your agreement.32 The deceptive manipulator uses persuasion rather than lies, but targets your lesser self with arguments, hoping you will not exercise your capacity to consider them. 33 The manipulators hope is that the victim will be less than fully the author of her action. Thus, it is the resemblance that this form of manipulation shares with deception that makes it morally problematic. To see the same point in a slightly different way, consider a question posed by Gideon Yaffe: If you are stuck with a companion who regularly gets you to do what she wants you to do, would you rather she were someone who gets her way by leading you to ignore those aims and wants with which you identify and do as she wants, or someone who
31 32

On the equivalence of lying and reticence as modes of deceiving, see Langton, Duty and Desolation. The bypass view of manipulation borrows from Alfred Mele, who describes ideal rational agents as capable of bringing their desires and pro-attitudes into alignment with their evaluative judgments about what they should do. Alfred Mele, Autonomous Agents (Oxford: Oxford University Press, 1995): 166-167. 33 Feinburg, Harmless Wrongdoing (Oxford: Oxford University Press, 1988): 201-202.

perniciously works on you to alter what you identify with so that it conforms with what she wants you to care about?34 Yaffe takes it as obvious that the second is worse the first. The reason is that the second leaves a lasting effect that the first avoids. When you discover that you have been led to satisfy anothers aims, you might rightly think of yourself as having been used. But when you realize that what you care about has been manipulated by someone else, you might worry that your self had changed in a way that you had not authorized. Who you are had been partly decided by someone who was not you. This sounds like a serious worry indeed, but as far as I can see, the only thing that makes it frightening is that we are explicitly told that our companions aims are pernicious. You can imagine a companion who works on you to alter what you identify with in a completely respectful way. He might, for instance, present you with reasons that he sincerely believes. He might introduce you to aspects of what he cares about that overlap already with things that you identify with, increasing the odds that you will enjoy it as well. The sheer fact of time spent together might mold you into holding similar values. And so on. Such give and take does change what we identify with, but it does so within the normal course of ordinary human relationships. It is only the fact that your companion in Yaffes example does this in a pernicious way that raises red flags. It is the possibility that he distorts your values in a calculating way that is meant to keep you from realizing it, or even understanding that it is happening. He need not be overtly deceiving you in order to be pernicious, but he must be attempting to hide from you something about what he is up to. The difference between the respectful companion and the pernicious one is not that one changes who you are more than the other, but that one takes the reins of deciding who you are away from you while you are not looking. He displaces you as the author of what you will do, or who you will be.


Gideon Yaffe, Indoctrination, Coercion, and Freedom of the Will, Philosophy and Phenomenological Research LXVII:2 (2003): 335-356, p. 338.

4 Among the examples presented so far by philosophers to illustrate what is the matter with manipulation, some attach modest or surreptitious penalties to certain courses of action. Others seek to intentionally impair the deliberative standpoint from which their victim will decide what to do. Both kinds of manipulation can be explained by a close similarity either to coercion or to deception. The fact that all of these cases fall into one category or the other might suggest that if one is careful to avoid anything resembling coercion or deception, then nothing is wrong with manipulation. It is just a part of life. If that is right, then Emily is justified in telling Lorelai to calm down. Besides the waste of a good afternoon, what is so bad about a golf date between grandfather and granddaughter? Lorelai believes that what is wrong is that Rory could not say no to her grandmothers invitationthat she was forced to do it. But as noted, there is neither trace of coercion nor anything deceptive about Emilys invitation. So what could the source of its normative forcefulness possibly be? Resources found in the reasons forbidding coercion or deception will help to answer this question. Out of respect for their autonomy, we owe it to other persons not to coerce or deceive them. Respect, as it has been described here, is a second-personal attitude. Reasons of respect arise from another persons authoritative standing, rather than from a state of the world. By hypothesis, any person with normal rational capacities has an equal claim to being respected. Put differently, the authority or standing of a person to demand respect for her autonomy does not depend on who she is, or what her relationship is with other people. Even if two people have never met before, they still owe each other this kind of respect. Any person

would have a standing to complain if we coerced or deceived her. In other words, the relationship of respect is completely general.35 To be clear, three features of the reasons that follow from autonomy are significant: their structure, their content (or the set of claims autonomy authorizes a person to make), and the set of persons to whom they can be appropriately addressed (call this their extension). First, their structure is second-personal. Reasons of autonomy follow from our recognition of anothers standing or authority to make claims against us, rather than our appraisal of her attributes. Second, their content is limited to claims that could be made against anyone, such as those warranted by the authorship principle. Although I have not said anything about what is excluded from their content, we can assume that only a very few claims against others would be warranted by autonomy alone. (The reasons why will shortly be clearer). Third, their extension is universal. As Kant imagined, we share a realm of reasons with all other autonomous beings. Autonomy alone gives an agent standing to issue reasons requiring respect. Now consider the possibility of reasons that are structurally similar to those provided by autonomy, but differ in content and extension. Take the attitude of love, for example. Love does not follow from an evaluation of attributes as sufficient to merit love, just as respect does not follow from an evaluation of attributes.36 We ought to respect another persons capacity to rationally guide his own life, regardless of whether he is good at it or not. Likewise we do not love those we love because they are excellent, lest the possibility that someone more excellent should come along would require us to love that person even more. In Darwalls terms, respect and love are both about recognition rather than appraisal. Like respect, love places another person in a position of practical authority in ones life. And like the reasons of respect, reasons of love are second-personal. They differ in that they are contingent rather than general. We are

Relationship is here used loosely, as we would seldom say that two people who have never met each other share a relationship. All that is meant is that a person can claim a right to be treated with respect, even by people she does not know. 36 My discussion here is informed by David Vellemans important article, Love as a Moral Emotion, in his Self to Self (Cambridge: Cambridge University Press, 2006): 70-109.

not morally required to love everyone in the same way that we are morally required to respect everyone. What follows has implications for the kinds of reasons love can provide to us. Very generally, respect requires that we leave people alone. When we respect another, our acknowledgment of her rational autonomy demands that we let her guide her own actions by her own choices. Respect thus checks our self-interested motives toward others whom we see as autonomous like ourselves.37 Respect accordingly entitles anyone to issue reasons to us to refrain from interference with their choices. Darwalls favorite example to differentiate second-personal from third-personal reasons involves discovering that someone is standing on your foot. Just in virtue of a relationship of respect, you have the authority to demand that the person step off of your foot. The reason for stepping aside is not that the world will be better if your foot is not under another, but that you asked that the person remove his foot. Reasons of love, I will suggest, are similar in that they follow from the authoritative standing of the beloved to issue reasons. They differ, however, in that someone with whom we are in a loving relationship will have a much greater standing in our lives to issue reasons than someone with whom we share only a relationship of respect. In relationships of love, we are not just checked by our respect in interacting with the other, but we feel safe to let ourselves be vulnerable to him.38 Vulnerability attends relationships of love because we allow the beloved to have a standing in our lives to issue reasons to us authoritatively in a domain that extends well beyond that of respect. We are made vulnerable because the reasons that we have for action will, in an important sense, no longer be entirely up to us. In between the limited and general second-personal authority of respect and the expansive and specific second-personal authority of love, there are a number of possible relationships in which parties have a standing to issue more reasons to each other than if they

37 38

J. David Velleman, Beyond Price, Ethics 118 (2008): 191-212, p. 201. Ibid., p. 202.

were strangers, but fewer reasons than if they were lovers.39 Consider the authority that friends have over each others reasons. If we are friends and I ask you to read my paper, you will have a reason to read it that you would not have if we were strangers. I cannot demand that you read my paper in the same way I can demand that you get off of my foot; it is not something you are obligated to do in virtue of respect for me as an autonomous person. However, within a relationship as friends, it is something that I can ask you to do. When pressed about your reason to read my paper, you might respond that I had asked you to, and my asking you gave you a reason. Of course, nothing would require you to answer in this way. If you wanted to resist the idea that reasons of friendship were second-personal, you might say that your reason for reading my paper was that you were in a position to comment on the paper in a way that would improve it substantially, and that doing so would be very good for me and not too much trouble for you. On balance, the world would be third-personally better if you read the paper than if you did not. While it might be possible to say that your reason to read my paper is therefore thirdpersonal, I think that many reasons of friendship resist such reinterpretation. We are often willing to work for the success of our friends projects when we are puzzled why anyone would adopt those projects, or even if we view them as intrinsically worthless. At such times, our reasons for helping our friends are simply that they have a standing in our lives to ask for our helpa standing that they would not occupy in the absence of the relationship.

5 Hopefully it will now be easier to see how Emily could get her granddaughter and her husband to go golfing together even if neither of them wanted to. As the wife of her husband and the grandmother of her granddaughter, Emily commands a post of considerable second-personal


For a general account of robust reason giving, see David Enoch, Giving Practical Reasons, Philosophers Imprint 11:4 (2011): 1-22.

authority in their lives. Given her standing to issue reasons within loving relationships, it is no wonder that her requests forcefulness was felt more vividly than it might have been otherwise. We can learn something about Emilys practical authority by recalling the authority provided by deontic autonomy.40 Another agents deontic autonomy creates second-personal reasons. When we recognize another agent as autonomous, we also recognize her standing to provide us with reasons for action. The content of these reasons is given by her demands upon us (or by her disposition to make these demands), combined with the fact that she has standing or authority to make the demands in question. The authorship principle tells us something about what kinds of demands an autonomous agent has standing to make. If I am autonomous, I can demand that others not interfere with my actions. Those to whom I address this demand will have reason to comply, the content of which will be given by my demand to them, given my standing to appropriately make this demand. This is distinguished from thinking that the reason is given by the outcomes of complying (or not complying) with the demand. The autonomy of others requires that we treat them with respect, no matter who they are. Regardless of their relationship to us, they have this minimal authority to address secondpersonal reasons to us. This, itself, creates a minimal relationship of sorts between us, which I have called a relationship of respect. The reasons that Emily gives to Rory and Richard are structurally similar. She has a standing or authority to address claims to them. This authority differs from that conferred by deontic autonomy, in two ways: first, it is something she has in virtue of her specific relationships with Rory and Richard; second, it allows her to address reasons to Rory and Richard in a much wider range of areas than deontic autonomy alone would allow. While respect requires that we leave others more or less to themselves, love allows us to ask a great deal of others (and for them to ask a great deal of us). These differences notwithstanding, there is also an important similarity. The content of the reason that Richard and Rory have to go golfing together is given by Emilys request (or her disposition to

I use the terms practical authority and second-personal authority interchangeably.

make such a request), given her standing to appropriately make such a request. It is just that in this case, her standing is not provided by her autonomy, but by her loving relationship with Richard and Emily. If this is right, it shows that persons can give each other reasons for action in virtue of certain kinds of shared relationships. Here are three abstract features of the phenomenon that I have in mind (where A and B are parties to the relationship and the is the action considered). (1) A intentionally uses her shared personal commitments with B to issue to B a reason to . (2) In the absence of As intentional action, B would not have rationally desired to . (3) A does not threaten B or misrepresent -ing or attempt to bypass Bs rational deliberation about -ing. What remains to be explained is when this type of reason-giving is manipulative, or wrong. There is no reason to think that all uses of practical authority within relationships of commitment are manipulative. However, there is at least a suspicion that in Emilys case, she uses her practical authority in a manipulative way. In the remainder of this section, I will consider this question. According to one view of manipulation, an action is manipulative if it intentionally attempts to get another person to do something that she would not otherwise want to do. Thomas Hill offers a view of this kind: Manipulation, broadly conceived, can perhaps be understood as intentionally causing or encouraging people to make the decisions one wants them to make by actively promoting their making the decisions in ways that rational persons would not want to make their decisions.41


Hill, Autonomy and Benevolent Lies, in Autonomy and Self-Respect (Cambridge: Cambridge University Press, 1991): 25-42, p. 33.

This account makes sense of why Lorelai accuses Emily of manipulation. Emily intentionally tries to bring it about that others make the decisions she wants them to make. Consider the following general formulation. M: A manipulates B if A intentionally causes B to act in a way that is in As

interest, but without coercion or deception. According to M, Emily does attempt to manipulate her husband and granddaughter. However, M is inadequate both as an interpretation of Hills account, and as a general statement of manipulation. For all we know, Bs action may be in Bs interest as well. We would not think that A had manipulated B ifsayA offered B a deal that was too good to refuse. If everyone Emily had invited was independently delighted by the prospect of a day on the golf course, it would be strange to charge her with manipulation. Moreover, it must be important to consider the second half of what Hill says, focusing not just on causing another person to act as you want her to, but doing so in a way that she would not otherwise want to make [her] decisions. M might be modified: M*: A manipulates B if A intentionally causes (without coercion or deception) B to act in a way that is in As interest, but is not independently in Bs interest. This formulation is still too expansive. Our commitments to other persons often provide us with reason to do things that we would not have otherwise wanted to do, but we seldom take this to be cause for resentment. Instead, we want those we love to demand things of usto change what we would otherwise do. When a 19th century Christian lyricist pleads more holiness give me, what he hopes is to be required to do more: more fit for the kingdom, more useful Id be.42 The Christian takes his commitment to God to purified by his willingness to be used for purposes that are not his own, but the idea that our relationships should change our aims is by no means a uniquely religious one. Our devotion to commitments and persons is indicated (or even partly constituted) by our alacrity for being taken away from ordinary pursuits.

Music and words by Philip T. Bliss, More Holiness Give Me, Sunshine for Sunday Schools (1873).

Sacrifices for our commitments are not made from manipulation. Rather, we would take it as cause for concern if our close friends didnt ever ask anything of us that we did not independently want to do anyway.43 Reluctance to make such demands would betray hesitancy about the friendship; a friendship whose members did not modify their behavior even slightly from what they would independently desire could scarcely be called a friendship at all. Friendship is so closely connected to a willingness to act contrary to our independent interests that sometimes the overture of another to perform a task for us that he would not otherwise desire is best declined if we want to prevent our association from becoming one of more significant commitment. It is politeness that requires such refusals be motivated by politeness. Another problem with M* is that it limits cases of As manipulation of B to those in which A causes B to act in As interest. One can imagine A trying to change the action that B will perform, but not for any reason relating to As interest. Instead, A might be motivated by a paternalistic desire for Bs action to serve Bs interests better than it otherwise would. Deception and coercion can both be used paternalistically, so it is natural to think that manipulation could be paternalistic as well. Deciding whose interest is being served should not be built into our understanding of manipulation, although it may sometimes be relevant to our moral assessment of particular cases. With these considerations in mind, M* might be revised again by suggesting that A manipulates B by getting B to act in a way that B would not have preferred if B had better understood the facts. This might be closest to what Hill has in mind, and it does seem to be morally concerning. If I alter the context in which you act such that you act in a way you would have not wanted to if only you knew more about your situation, you would likely resent my manipulation of your deliberative environment. However, this just returns focus to what I


Cf. Samuel Scheffler, Projects, Relationships, and Reasons, in Wallace, Pettit, Scheffler & Smith (eds.) Reason and Value (Oxford: Oxford University Press, 2004): 247-269.

described above as deceptive manipulation, since it would involve getting the manipulee to act without understanding the relevant reasons. Could it be, then, that manipulation is never wrong except when combined with elements of coercion or deception? The normative power that people in relationships of commitment can wield (or at least, utilize) over each other makes me think that there must be some set of norms that govern such relationships beyond those forbidding coercion and deception. One candidate for such a norm on the use of practical authority within a relationship of commitment might be to refrain from making requests that are over-demanding or burdensome. Perhaps Emilys mistake is that she asks too much of her family members, and this makes her manipulative. Unfortunately, this way of specifying the norm does not look promising. It seems unlikely that there are any general facts about what will count as burdensome or overdemanding. The extent to which a use of practical authority within a given relationship is appropriate depends on particular features of that relationship. What authority a relationship of commitment confers can be decided only by those within it. What would be an over-demanding request in one relationship might be appropriate in another. It would be impossible to demarcate a general threshold of demandingness beyond which requests were inappropriate. A better way of specifying the norm would recognize that what will count as manipulative will vary from one relationship to another. One way of capturing this sensitivity to detail would be to state the norm as one of reciprocity within a relationship, perhaps in this way: As request that B perform some action () is manipulative if A would resent Bs request that A perform an action of roughly equal seriousness as -ing. I will call this the norm of symmetry. The thought behind this formulation is that persons within relationships of commitment often understand membership to confer an equal practical authority to each member. In other words, if we are friends, it is natural to think that the friendship will give me roughly equal practical authority in your life as it gives you in mine. It would be manipulative of me to ask a great favor of you as my friend, if at the same time I would not

approve of your using our friendship to ask a comparable favor of me. A friendship that extended deeply asymmetrical practical authority to its members might be thought in some way unhealthy. For many relationships of commitment, a norm of this type seems fitting. It makes sense of why parties to personal relationships will sometimes investigate whether a particular request is reasonable by inquiring about whether a similar request made in the opposite direction would be accepted by the person doing the asking. In effect, it is a redeployment of the basic moral question: How would you feel if I did that to you?44 The trouble with symmetry as a specification of norm against manipulation is that for many relationships, parties to the relationship do not seem to have equal practical authority in each others lives, and this does not appear to be a problem. First, consider cases in which asymmetry seems acceptable. For example, a priest may have a valuable personal relationship with a parishioner, and both may agree that this gives the other a kind of practical authority. However, they may also agree that the priest has greater practical authority in life of the parishioner than the parishioner does in the life of the priest (or perhaps the opposite is true). In either case, the symmetry norm would suggest that the relationship was inappropriately manipulative. But this seems mistaken. Both priest and parishioner might want their relationship to be one of asymmetrical practical authority, and there is no obvious reason to think them mistaken. Alternatively, just as there are cases where asymmetry may be appropriate, there may also be cases where symmetry is not appropriate. Consider two friends, one sensitive, the other unflappable.45 The unflappable friend might make a request of her more sensitive counterpart, thinking to herself, I wouldnt mind at all if my friend were to make a comparable request of me, so I should not be concerned in making such a request to my friend. Her reason for this might be that if her friend were to make such a request to her, the reason provided by

44 45

Cf. Nagel, The Possibility of Altruism (Princeton: Princeton University Press, 1970). I am grateful to XXXX for this example.

the request would not easily override other things that she cared about. As an unflappable person she would be able to weigh the request against reasons for other actions without feeling that she had to comply. However, this might not be true of someone with a different disposition. Her sensitive friend might feel compelled to comply with a request that she would feel compelled to comply with. This might lead to reasonable cause for concern. If one friend routinely asked favors of another that the latter felt always compelled to comply with, this might begin to look manipulative even if formally compatible with symmetry in practical authority. If symmetry is an inadequate specification, consider a still more general alternative. Here is a final account of some of the elements of manipulation. A manipulates B if: (1) A intentionally uses her shared personal commitments with B to issue to B a reason to . (2) In the absence of As intentional action, B would not have rationally desired to . (3) A does not threaten B or misrepresent -ing or attempt to bypass Bs rational deliberation about -ing. (4) As use of personal commitments with B involves an abuse of the relationship, in that there is no shared understanding about what practical authority the relationship confers with which it is compatible. Conditions (1)-(3) are familiar already. They pick out how persons within a certain class of relationships can give each other reasons for action. Condition (4) replaces symmetry as the norm showing when the use of a relationship is manipulative. Call this account relational manipulation, to separate it from its coercion-like and deception-like counterparts. On this account, what places one person in a position that allows manipulation of another is a relationship of commitment, understood as a shared standing to issue second-personal reasons exceeding those of respect. In all such relationships, parties to the relationship give each other reasons that (intentionally) change what reasons the other has

to act on. The shared relationship gives persons within it the authority to issue these reasons. Manipulation is the abuse of that authority. Condition (4) attempts to say something quite general about when ones authority within a relationship is abused. It holds that the phenomenon of reason-giving pointed out in conditions (1)-(3) becomes manipulative when there is no shared understanding of the relationship that licenses it. Developing a shared understanding of a relationship involves recognizing some uses of the practical authority conferred by the relationship as appropriate, and others as inappropriate. This process need not be deliberate, or even conscious. Shared understandings of what relationships do and do not allow us to ask of others can be built up gradually while being recognized implicitly. We may only realize that a given request is not in keeping with the shared understandings of a relationship when we imagine asking it. When I contemplate asking a friend to accompany me to a conference, I may be struck by the thought that asking that particular person to spend an entire weekend keeping me company would be beyond what is in keeping with our relationship. Given certain shared understandings, this might simply be asking too much, notwithstanding that the request would provide my friend with reasons to agree. Finally we have the tools to see what is manipulative about Emilys invitation to Rory. An offer from Emily to go golfing is not like an offer from a classmate, or a sports coach. Were she to be passed a registration form during class, Rory would have only reasons of her own to consider. Did she like golfing? Would it be worth the time away from other pursuits? etc. The fact that it is EmilyRorys grandmotherwho extends the invitation, changes the reasons relevant to the decision. In asking Rory to go golfing, she also gives her a reason to go. The content of this reason is given by Emilys request, given her authoritative standing in Rorys life.46


Some counter that the reason actually follows from the fact that Emilys feelings would be hurt if Rory did not accept. In the terms used here, this is to suggest the reason is third- rather than second-personal.

So far, nothing is wrong with the request, although we can see why such requests should be made with care. We can sometimes induce those in close relationships with us to change what they will do just by making an offer.47 As Lorelai points out to Emily, she might be making an offer to Rory that, given her position of special second-personal authority, Rory cannot refuse. This is not to say it is a request of any special severity. What is manipulative is not the request itself, but the request in context of the relationship. One possibility is that the relationship includes a shared understanding according to which Emily, as Rorys grandmother, can recommend certain activities and associations. Another is that the relationship includes a shared understanding according to which Emily cannot appropriately use her practical authority in Rorys life in this way, but should instead keep a respectful distance. A third possibility is that there is no shared understanding about whether this is an appropriate use of the relationship. Lorelais criticism might be understood as based on the complaint that Emilys request went beyond the level of involvement sanctioned by the shared understandings of their relationship. Lorelai understood their relationship as one in which Emily ought to have kept a respectful distance from her granddaughters decisions. And while Emily might not have shared that understanding, this at least shows that there was no shared understanding with which Emilys actions were compatible. Even when there is no shared understanding of what a relationship allows, parties to the relationship may mistakenly believe such an understanding obtains. They may also be mistaken about the content of shared understandings. One member of a friendship might think
I doubt this, for two reasons. First, it seems to me that the reason would persist even if Rory had assurances that Emilys welfare would not be set back by her refusal. Suppose, for example, that Rory had a twin; that Rory knew that Emily would be satisfied so long as one of her grandchildren went on the trip; and that she further knew that her twin would go if she declined. I still think that Emilys request would provide her with a reason to go. Second, the fact that Rory is concerned about the prospect of Emilys feelings being hurt does not reveal her concern to be third-personal. It is natural to be concerned about the feelings of those who have an authoritative standing in our lives due to non-instrumentally valuable relationships. At worst (for my account), this phenomenon is indeterminate between explaining Emilys action as coercive manipulation or as relational manipulation. 47 In her perceptive discussion of the problem, Onora ONeill observes, in intimate relationships it is all too easy to make the other an offer he or she cannot refuse. See her Between Consenting Adults, in Constructions of Reason (Cambridge: Cambridge University Press, 1989): 105-125, p. 120.

the relationship is one where both members can appropriately ask much of each other, while the other might believe that each will ask for help only for important or urgent reasons. Misunderstandings of this kind are common in any relationship of commitment. It would be strange to think that whenever they occurred, someone was being manipulative. This suggests one clarification. If a party to a relationship mistakenly believes there to be a shared understanding licensing a particular action, and is not morally responsible for this false belief, then taking the action in question should not be judged manipulative. For example, if I reasonably (but falsely) believed that our friendship allowed me to appropriately ask my friend to accompany me to the conference, then my request is not manipulative, but merely ignorant or confused. In the case of Emily, Lorelai clearly thinks that Emily is aware that there is no shared understanding of the relationship according to which her involvement is appropriate. That Emily involves herself anyway reveals her designs are manipulative. One noteworthy difference with other basic kinds of wrongdoing is that relational manipulation does not appear to violate the authorship principle. Manipulating a person with whom one shares a relationship does nothing to deny her standing as the author of her action. This is because relational manipulation does nothing to impair the persons deliberation about her options, or to modify her options. What manipulation does do is alter the reasons that the person has for acting. In addition to whatever reasons there were before the manipulative request was made, the request adds a set of second-personal considerations that count in favor of complying with the request. This is why it can sometimes make sense to allow yourself to be manipulated, even after you get a clear-headed understanding of the situation you are in. It is in this sense, I believe, that manipulation counts as a kind of forcing. We can be forced to help hapless neighbors with yard work, or to attend pointless meetings run by a close friend, but not because they will do something to us if we dont. Rather, we are forced by the fact that this is

what the balance of our practical reasons recommends.48 Still, we may still be doing something that is bad for us, despite the fact that we have reason to do it. Requests within relationships of commitments should be made with care. Ultimately, the account I have sketched here is motivated by a fairly simple ideal of interpersonal relationships. The authority we have in the lives of those close to us is often significant. It is tempting to think that the ideal we should strive for is one in which we refrain from using this authority for our own ends. But this is implausible. Human relationships are naturally used to advance the ends of their membersa fact that, alone, poses no moral worry. Complications arise when this authority is used too much. There may be no general way of explaining how much is too much, since all relationships are different. Some friendships are more intimate, and others are more respectful. The relationships among neighbors in the libertarian West may differ from those had by members of a commune. I see no reason to think that there is some fact of the matter about how much authority a given relationship should confer. The important thing is that we should not demand more of friends and loved ones than the specific facts of our particular relationships allow. Strangers should treat each other with respect; intimates should be sensitive to each others understanding of their relationship. To do otherwise would be to use the relationship for reasons the other cannot share. It would be to make an exception for oneself to the norms of the relationship.

Conclusion It is sometimes taken to be a peculiar fact about relationships that persons who are close to us-and care about us a great deal--can also hurt us more than strangers who dont care about our welfare at all. Stranger still, they can do this without meaning to. I have tried to give an account of the reasons that arise from relationships that seeks to understand these facts. We can wrong


For a discussion on the relationship between practical reasons and freedom, see Nomy Arpaly, Merit, Meaning, and Human Bondage (Princeton: Princeton University Press, 2006): Ch. 4.

others by coercing and deceiving them, but we can also do wrong by manipulating them. However, we cannot manipulate just anyone, at least in the way that I have had in mind here. Instead, we are most vulnerable to manipulation from, and most likely to manipulate, those to whom we are closest. Further, whether a given question or offer counts as manipulative depends on facts beyond those of just that inquiry or offer. The nature of give and take in the relationship has to be taken into account when testing whether an offer is appropriate, and that can be a more subtle and challenging matter than when only one action is under moral scrutiny. But the news is not all bad. Relationships create new moral risks, but how could they be morally rewarding if they didnt?