Social Science - Philosophy Assignment 272

Manipulation: Theory and Practice

Christian Coons (ed.), Michael Weber (ed.)

https://doi.org/10.1093/acprof:oso/9780199338207.001.0001

Published: 2014 Online ISBN: 9780190228446 Print ISBN: 9780199338207

Search in this book

CHAPTER

https://doi.org/10.1093/acprof:oso/9780199338207.003.0004 Pages 73–97

Published: August 2014

Abstract

Keywords: manipulation, rational capacities, autonomy, agency, deception, undue influence, reasons

Subject: Social and Political Philosophy, Moral Philosophy

Collection: Oxford Scholarship Online

3 Towards a Theory of Interpersonal Manipulation  Moti Gorin

This chapter begins to chart out a novel account of interpersonal manipulation. Most of the arguments

are negative, as the bulk of the essay is devoted to showing what manipulation is not. After criticizing

several more or less plausible claims about the nature of manipulation—that manipulation necessarily

involves deception, harm, the undermining of autonomy, or the bypassing or subversion of the

manipulated agent’s rational capacities—the chapter sketches a more promising view. It concludes

with some very brief remarks on how the theory of manipulation that is begun to be developed here

might help provide the basis for a new and interesting ethical analysis of manipulation.

Introduction

People are complicated beings, exhibiting an extremely wide range of behaviors that are due to an equally

wide variety of causes. Consequently, there are myriad means available to in�uence this behavior. We make

claims, both true and false. We construct good arguments and bad ones. We make di�erent sounds and

facial expressions. We clothe, decorate, situate, and move our bodies in seemingly in�nite ways. We make

use of tools and other sorts of artifacts. We alter our environment and in so doing stimulate our perceptual,

cognitive, and emotional faculties. Each of these means of interpersonal in�uence can be used

manipulatively, though none of them is essentially manipulative.

In what follows, I begin to chart a novel account of interpersonal manipulation. Most of the arguments are

negative, as the bulk of the essay is devoted to showing what manipulation is not. After criticizing several

more or less plausible accounts of manipulation I sketch what I take to be a more promising view. I conclude

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

Manipulation and Deception

with some very brief remarks on how the theory of manipulation I begin to develop here might help provide

the basis for a new and interesting ethical analysis of manipulation.

Manipulation and Common Wrongs

Manipulation commonly involves ethically suspect behavior such as deceiving, harming, undermining

autonomy, or bypassing or subverting the rational capacities. Hence, it is tempting to think there is some

necessary connection between manipulation and these other things. There are also theoretical

advantages to insisting on tight links between them, for though deception, harm, autonomy, and the

rational capacities remain to varying degrees contested concepts it is at least fairly clear what the major

competing normative ethical theories have to say about them. A necessary connection between one or more

of these concepts and manipulation would allow for the derivation of conclusions regarding the nature of

manipulation from claims about deception, harm, autonomy, or the bypassing or subverting of the rational

capacities. The most interesting ethical questions about manipulation would turn out to be questions about

other phenomena whose natures have been more frequently discussed and which are better understood. For

example, if manipulation always involved deception, then answers to questions about the ethical status of

deception would also serve as answers to questions about the ethical status of manipulation. This would

leave us with a relatively tidy way to approach questions about the normative dimension of manipulation.

p. 74

In the following four sections I examine the relationships between manipulation and deception,

manipulation and harm, manipulation and autonomy, and manipulation and the rational capacities. I argue

that although manipulation often does involve one or more of these, it does not always do so. An account of

manipulation that reduces its normative signi�cance to concerns raised by deception, harm, and threats to

autonomy or the rational capacities will fail to capture much that is interesting and important about

manipulation. Such an account will therefore remain incomplete. I will begin with a discussion of the

relationship between manipulation and deception and then move on to discuss harm, autonomy, and the

rational capacities. My strategy will be to motivate accounts of manipulation according to which these

wrong-making features are necessary conditions of manipulation and then to provide counter-examples to

these accounts. The central conclusion of this section is that manipulation does not essentially involve

deception, harm, the undermining of autonomy, or the bypassing or subverting of the rational capacities.

None of these can provide a necessary condition in the analysis of interpersonal manipulation.

The �rst account I will examine pays special attention to the epistemic features of manipulative interactions

and in particular to the role of deception in these interactions. On this account, which I will call the

Deception-Based View, manipulation always involves some element of deception. A defender of this view

can correctly point out that many paradigmatic cases of manipulation involve deception and that deception

may enter into a manipulative encounter in more than one way. First and most crudely, a manipulator

may lie—that is, he may state something he knows to be false with the intention that it be believed to be

true. Here is one example of this.

p. 75

Not Credible: Henry wishes to undermine the credibility of his colleague Elizabeth. He lies to her

about various matters on which she rightly takes him to be an authority. Later, when Elizabeth is

having a conversation with other experts in Henry’s �eld, she relies on the “information” Henry

provided her. The specialists, who correctly judge that Elizabeth is advancing false claims, begin to

doubt her competence. The experts’ judgments that Elizabeth is an unreliable source of

information or that she is incompetent are products of Henry’s manipulation.

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

In this case of epistemic manipulation, Henry has manipulated Elizabeth as well as his peers, and his

method of doing so included the telling of lies as its central component.

Less crudely, a manipulator may say something that is true but which he intends will lead his interlocutor to

believe something false. Depending on the other beliefs an agent has and on the context of the exchange, the

acceptance of a true belief may lead to her acceptance of a false belief. Here is one such case.

Synagogue: David is romantically interested in Susan and so is his friend Jack. David knows Jack is a

committed Catholic who prefers to date other Catholics. David knows that Susan, too, is Catholic

but he does not wish Jack to know this, as David would like to reduce the amount of competition he

might face for Susan’s a�ection. David recently saw Susan entering a synagogue. Though he knows

Susan was there only to meet with the rabbi about an upcoming fundraiser for a

nondenominational charity, the next time he has lunch with Jack he mentions that he saw Susan at

the synagogue. David intends that this will lead Jack to believe that Susan is Jewish and,

consequently, that Jack will come to believe that Susan is not a viable romantic option for him.

David states something he believes to be true and he intends that Jack accept the statement as being true.

Nevertheless, David intends that Jack’s acceptance of a true claim will lead to his holding a false belief and

ultimately that this will lead to the behavior David is seeking from Jack. David’s behavior is both

manipulative and deceptive but it does not involve a lie.

The Deception-Based View of manipulation captures an important feature of manipulation, namely that it

can “prevent [a manipulee] from governing herself with an accurate understanding of her situation.” In the

cases discussed so far, manipulators do this by causing manipulees to have false beliefs whose content

extends beyond the intentions of the manipulator, though of course the manipulators also deceive the

manipulees about their intentions (otherwise it would not be easy to deceive them about anything else). But

manipulators sometimes prevent manipulees from having an accurate understanding of their situation by

causing them to have false beliefs or to fail to have salient true beliefs whose content is limited to the ends

at which the manipulators’ actions are aimed and the role the manipulees play in the achievement of those

ends. In such cases, the manipulators’ intentions are “masked” though the manipulees are not deceived

about anything external to the intentions of the manipulators. Here is such a case.

p. 76 1

Flattery: Carlos approaches his boss Lucinda at the company holiday party and tells her that her

recent restructuring of the company’s distribution system was altogether brilliant. Though Carlos

happens to believe Lucinda’s recent performance really was brilliant, he would have told her this

even if he believed her e�orts displayed rank incompetence. Carlos knows he is telling his boss

something she has heard from many others and which she already believes, and he believes that

due to his own limited business experience Lucinda probably will not take his opinion to carry

much weight as an evaluation of her work. Carlos believes the only value of his expressing his

opinion lies in its potentially causing Lucinda to be positively disposed towards him, and he wants

badly for her to be so disposed in light of his recent performance review, during which Lucinda

expressed serious concerns about Carlos’s ability meet the requirements of his job. Carlos is

motivated to appear to compliment Lucinda exclusively by the e�ect he thinks doing so may have

on her attitudes towards him.

Carlos does not deceive Lucinda about his opinions of her work but he does act deceptively insofar as he

wants it to appear to Lucinda that his comment was motivated by his beliefs regarding the features of

Lucinda’s behavior that really do justify a compliment, and not exclusively by his desire to get into her good

graces. Carlos must rightly assume that if Lucinda believed he was merely trying to ingratiate himself to her,

his action would be unlikely to elicit attitudes that would bene�t him. By masking his intentions with

respect to Lucinda’s attitudes towards him Carlos attempts to mislead Lucinda about the purpose of his

p. 77

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

disclosing (what just happens to be) his opinion to Lucinda. The masking of his intentions is necessary for

their satisfaction and is a central element in his plan. Carlos is attempting to “prevent [Lucinda] from

governing herself with an accurate understanding of her situation” insofar as the success of his plan—that

is, that Lucinda have certain attitudes about him—depends on her misconstruing the purpose of their

interaction. Carlos acts deceptively and manipulatively, though the scope of his deception is limited to the

content of his intentions.

In all cases of successful deception the intentions of the deceiver will to some extent remain hidden. In most

cases of deception the masking of the intentions is of derivative, instrumental importance from the point of

view of the deceiver, as the more central aim of the deceiver is the acceptance by the deceived of false beliefs

about some state of a�airs that is independent of the intentions that lie behind the act of deception. But in

other cases of deception the object of the deception just is the content of the deceiver’s intentions. The

victim of the deception comes to have false beliefs only about what the deceiver is doing in interacting with

her. As the case of Carlos and Lucinda illustrates, it is possible for an agent to speak the truth while

nevertheless dissembling, as the content of the propositions asserted (e.g., that Lucinda’s performance was

brilliant) is independent of the content of the intentions that underlie their assertion (e.g., that Lucinda

come to view Carlos in a more favorable light.)

When one agent interacts with another agent the latter typically will have expectations about the intentions

of the former and the role she (the latter) plays in those intentions. Generally these expectations are not the

product of any explicit statement or agreement but are, rather, assumed to underlie the interaction. In

typical cases of communication an agent expects that her communicative partner adheres to certain norms

of discourse—for example, that she be neither more nor less informative than necessary, that she speaks

with the intention to convey what she believes to be true, that she says only what is relevant, and that she is

reasonably careful to avoid saying things that may lead to misconceptions or confusion.2

I propose to add to this list a Transparency Norm, which requires that an interactive partner not hide her

intentions in interacting when these intentions are relevant to the intentions or interests of the person with

whom she is interacting. Unlike the truth-telling norm, which is quite general and has application in most

(if not all) contexts, the Transparency Norm may have a more limited applicability, the criteria for which

will vary with context. For current purposes, I hope only to have shown how deceptive manipulation may

involve a particularly nuanced kind of deception, one in which a manipulee is deceived not about the truth

value of what the manipulator is claiming but, rather, about what both manipulator and (as a consequence)

manipulee are doing. Indeed, in all cases of deceptive manipulation, whether the content of the deception is

limited to the intentions of the manipulator or extends beyond them, a central aim of the manipulator is to

deceive the manipulee about the role the latter plays in the plans of the former. Unlike in non-manipulative

deception, where the point of the interaction is to cause false beliefs with content extending beyond the

intentions of the manipulator, in cases of manipulative deception such beliefs, if they are at all present, are

of derivative value to the manipulator, whose central concern is to mask her intentions and the role the

manipulee plays in these intentions. The Transparency Norm would rule out deceptive manipulation as well

as most standard cases of deception such as lying and is thus more general than a standard truth-telling

norm. It is by playing on the expectations of manipulees, expectations generated by adherence to the

Transparency Norm, that manipulators prevent manipulees from governing themselves with an accurate

understanding of their situation.

p. 78

In What We Owe to Each Other, Thomas Scanlon discusses how our causing others to have expectations about

our behavior can generate moral obligations. In this context, he articulates a principle meant to rule out

unjusti�ed manipulation. He calls this principle “Principle M,” and it requires that (in certain

circumstances) agents not hide their (relevant) intentions in interacting with others.

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

Principle M: In the absence of special justi�cation, it is not permissible for one person, A, in order to

get another person, B, to do some act, X (which A wants B to do and which B is morally free to do or

not do but would otherwise not do), to lead B to expect that if he or she does X then A will do

Y(which B wants but believes that A will otherwise not do), when in fact A has no intention of doing

Y if B does X, and A can reasonably foresee that B will su�er signi�cant loss if he or she does X

and A does not reciprocate by doing Y.

p. 79 3

According to Scanlon, Principle M is a valid moral principle. This is because:

[c] onsidering the matter from the point of view of potential victims of manipulation, there is a

strong generic reason to want to be able to direct one’s e�orts and resources toward aims one has

chosen and not to have one’s planning co-opted...whenever this suits someone else’s purposes.4

Here Scanlon voices a concern similar to that expressed by Buss when she says that manipulation can

“prevent [a manipulee] from governing herself with an accurate understanding of her situation.” The

explanation for Principle M—that is, that people have strong reasons to want to be able to direct their

energies towards aims they have chosen, and that hiding one’s intentions when interacting with others can

undermine this ability—may capture one ethically troubling element that is sometimes present when one

agent manipulates another. The basic idea seems to be that when one’s intentions impact the intentions of

others, it can be wrong to mislead others about what one’s intentions really are. Scanlon goes on to discuss

other more general but related principles that he thinks account for the wrongness of promise-breaking and

lying, and he claims that these principles are generalizations of Principle M. On his view, unjusti�ed

manipulation is a special case of lying, and thus Scanlon seems committed to the Deception-Based View of

manipulation.

5

6p. 80

In each of the cases discussed so far, a manipulator deceives a manipulee by making claims (whether true or

false). However, a manipulator may avoid making claims and yet use deception to control the behavior of

others. For example, advertisers frequently arrange non-propositional visual and auditory stimuli in ways

that associate the products they are trying to sell (or the policies they are trying to promote) with the

preferences of members of the target demographic, even when there is no rational or causal connection

between the stimuli and the products (or policies) with which they are being associated. Many such cases

will clearly count as manipulative. Or, a manipulator may make changes in the environment which are

intended to lead to the manipulee’s holding false beliefs and behaving on the basis of doing so. Carol Rovane

provides a nice example of this kind of manipulation:

[Y] ou are about to leave the house without your umbrella. And... I decide that it would be amusing

to get you to take it.... I happen to know that you always take your umbrella on days when your

housemates take theirs. I also happen to know that there is an umbrella stand near the door which

is usually full of umbrellas, except on days when your housemates have taken them. So I remove all

of the umbrellas but yours from the stand with the following aim: you will notice that the other

umbrellas are gone, you will infer that your housemates have taken their umbrellas, and you will

decide to follow suit by taking yours.7

Here the manipulator avoids making any claims at all and yet the manipulation is deceptive.

The cases presented above are representative of a large class of manipulative actions of the sort captured by

the Deception-Based View. However, there are counterexamples to the Deception-Based View.

O� the Wagon: Wilson and Adams are up for promotion, though only one of them will get the job.

Wilson is a recovering alcoholic and Adams sets out to encourage a relapse, intending this to

disqualify Wilson for the promotion. Adams consistently drinks alcohol in front of Wilson, o�ers

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

Manipulation and Harm

her alcoholic beverages, vividly describes to her whatever bene�ts there are to drinking and to

drunkenness, and so on, all the while making no secret of his intentions. During a moment of

weakness brought on by a particularly di�cult and stressful event Adams takes a drink, which

leads to more drinks, missed days at work, and an overall decreased ability to meet the demands of

her job. When the time comes to announce who will be promoted, Adams is told by her managers

that her recent poor performance has made it impossible for them to give her the new job and that

they have selected Wilson for the promotion.

p. 81

Wilson has manipulated Adams by engaging her compulsion to drink alcohol. And Adams’s awareness of

Wilson’s intentions does not undermine the intuition that this is a genuine case of manipulation. In this

case the manipulator does not deceive the manipulee about anything. The manipulator’s intentions are

known to the manipulee and no false claims are advanced. Therefore, manipulation need not involve any

deception. The Deception-Based View is false.

According to the next account of manipulation I will examine—the Harm-Based View—manipulation

essentially involves harm, and this is what provides us with a reason to avoid engaging in it. The Harm-

Based View accounts for the fact that often when we criticize an instance of manipulation one of the features

we single out is the harm done to the manipulee. It also accounts for the fact that manipulators often do

advance their own interests at the expense of those whom they manipulate. David in Synagogue seeks to

increase the likelihood of his getting what he wants (a relationship with Susan) by decreasing the likelihood

of Jack’s getting what he wants (also a relationship with Susan), and Adams in O� the Wagon improves his

situation by making Wilson signi�cantly worse o�. Scanlon’s Principle M involves one agent deliberately

gaining advantage at the expense of another agent who, as a result of their interaction, would su�er

signi�cant loss. Indeed, it might be thought that the motivation for the Deception-Based View is grounded

at a deeper level in a concern about harm. Perhaps a defender of the Deception-Based View mistakes the

importance of process (deception) with that of a salient consequence (harm) of that process. In any case, an

account of manipulation that takes harm to be an essential normatively relevant feature will capture some

cases of manipulation that are left out by the Deception-Based View—for example, O� the Wagon. It will also

explain why manipulation often does involve deception, for people who are mistaken about their situation

—for example, about the consequences of their behavior—are more likely to behave in ways that are

detrimental to their own interests.

p. 82

Typically, people resort to manipulating others when they believe other methods of in�uence will fail.

Sometimes there simply are no good reasons that can be given to someone to motivate her to behave in

some way—not because she is not amenable to reason but because she is amenable to reason and what is

being asked of her is contrary to reason. When an agent believes some possible action of hers will be

detrimental to her interests she probably will be strongly disposed to avoid doing that action. Moreover, if

she has su�cient evidence for her belief and is rational there may be no good argument to convince her

otherwise. In such cases, it may be necessary for the person seeking control to manipulate the agent into

doing whatever it is she wants her to do. As an e�ective means of directing people to do voluntarily what is

not in their best interest, at least according to their own considered judgment (which may or may not be

consistent with their judgment at the time of the manipulated act), manipulation often does involve harm to

the manipulated agent.

But the Harm-Based View does not stand up to much scrutiny. Perhaps the easiest way to see this is by

re�ecting on cases of manipulative paternalism. Though it is di�cult non-manipulatively to direct people

to act in ways that are inconsistent with their own considered judgments regarding their interests, people

are prone to acting against their interests on their own, sometimes consciously. Manipulation can be used to

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

Manipulation and Autonomy

prevent them from doing so. The “libertarian paternalist” policies proposed by Richard Thaler and Cass

Sunstein are intended to cause people to behave in ways that bene�t them and they do so in ways that

exploit irrational (or, weaker, nonrational) tendencies. For example, if a cafeteria manager gets people to

eat healthful foods by carefully arranging the order in which the food choices are displayed in the cafeteria,

it is plausible that he has manipulated his customers to act in ways that bene�t them. Here is a more

straightforward example.

8

9

Dementia: Mildred, who su�ers from dementia, appears to have an infection. Her son Nathaniel

wants her to go to the hospital but is unable to persuade her to do so by citing the reasons that

support her doing so (e.g., that infections left untreated may be life-threatening, that the hospital

is the best place to treat the infection, etc.). Nathan knows that his mother would go to the hospital

if she were told to do so by his father. The problem is, his father has been dead for a number of

years. However, due to her dementia, Mildred often mistakes her son for her husband. Nathaniel

waits until his mother calls him by his father’s name and then, pretending to be his father, tells her

that he would like her to go to the hospital to have her infection treated. She agrees.

p. 83

This case raises a number of di�cult ethical questions. However, it should be clear that Nathaniel has

manipulated his mother and also that he neither intended harm nor likely brought any about. His action was

manipulative but bene�cent. Unless we implausibly stipulate that to manipulate someone is ipso facto to

harm her, the Harm-Based View will be subject to many similar counter-examples.

The third view I will examine is the Autonomy-Undermining View of manipulation. According to this

account manipulation essentially involves the undermining of an agent’s autonomy. The Autonomy-

Undermining View is more di�cult to assess than the previous two accounts. Theories of autonomy vary

and thus an account of manipulation that makes autonomy-undermining central will need to specify which

notion of autonomy is at issue. Broadly speaking, there are two approaches one may take in analyzing

autonomy. The �rst is purely “internalist” in that it seeks to locate autonomy in the relations between an

agent’s propositional attitudes, irrespective of the source of those attitudes or the processes underlying

their acquisition and development. The second is “externalist” in that it looks to the sources of an agent’s

motivational set and the manner in which members of that set were acquired and arranged—that is, their

history. Externalist accounts may themselves di�er signi�cantly in how they distinguish between

autonomy-conducive histories and autonomy-undermining histories. In this section, I brie�y describe

internalist and externalist accounts of autonomy and then argue that whichever of these provides the best

theory of autonomy, each is consistent with manipulation. Manipulation does not entail the undermining of

autonomy.

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

Internalist Theories of Autonomy

Externalist Theories of Autonomy

p. 84

On one in�uential internalist account of autonomy all that matters is the degree of coherence between �rst-

order and higher-order propositional attitudes. An autonomy-undermining theory of manipulation that

construes autonomy in this way must insist that manipulators intervene between their manipulees’ �rst-

order attitudes and their higher-order attitudes. To illustrate, suppose an agent has a second-order desire

D2 that some �rst-order desire D1 of his not move him to action. According to the internalist, a manipulator

may undermine this agent’s autonomy by, say, altering the intensity of D1 so that D1 is now action-causing

for the agent. If the agent acts on D1 despite the presence of D2, then the agent has not acted autonomously.

Part of the explanation for this is that he was manipulated, since it is the manipulation that led to the

misalignment between the relevant attitudes. According to the internalist theory an action is autonomous

when higher-order and lower-order attitudes regarding that action cohere in a speci�c way, and thus for

manipulation to be essentially autonomy-undermining is for it to be essentially coherence-undermining.

10

The problem with trying to explain manipulation by reference to an internalist theory of autonomy is that

there are cases of manipulation that clearly do not threaten the coherence of the manipulated agent’s

attitudes. Drawing on the case provided in the last paragraph, a manipulator may leave D1 alone, opting

instead to alter D2 so that it coheres with D1. If the agent then acts on D1 he will have done so autonomously

according to the internalist. Similarly, a manipulator may alter attitudes on both higher and lower levels so

that an autonomous decision not to do X becomes an autonomous decision to do X.

For example, as a result of being exposed to subliminal messages an agent who wants to avoid hurting her

friend and also wants to want to avoid hurting her friend might form the desire to slap her friend as well as

the desire to be the kind of person who desires to slap her friend. If as a result of this she does slap her

friend, this would constitute a case of manipulation, though according to the internalist account of

autonomy the agent acted autonomously. On this picture manipulation cannot be essentially autonomy

undermining, for autonomy is preserved despite the manipulation or even as a result of it.

I do not believe manipulation necessarily undermines autonomy. However, in order to vindicate this claim I

will need to do more than merely rehearse some of the well-known objections to internalist theories of

autonomy. I will need to show how manipulation is consistent with autonomy as the latter is construed by

externalist theories as well.

p. 85

Before discussing any particular externalist theories of autonomy it is important to note an ambiguity about

what ‘external’ is supposed to denote in such theories. On the one hand, there are questions about the

sources from which and the processes by which an agent came to hold the propositional attitudes or, more

broadly, to be in the behavior-underlying states in question. On the other hand, there are questions about

the agent’s attitudes about those processes. I call theories that focus exclusively on the �rst class of

questions pure externalist theories. Such theories seek to distinguish between autonomous and non-

autonomous behavior (broadly construed to include the acquisition/holding of propositional attitudes,

emotional �uctuations, etc.) by reference to the processes that lead up to the states of the agent that

underlie the behavior. According to a pure externalist theory of autonomy, the truth of autonomy claims can

be determined in the absence of any reference to the agent’s attitudes about her own states or the processes

that lead up to them.

The second class of externalist theories, which I label mixed theories, hold that in answering the question of

whether or not some agent is autonomous with respect to some behavior we must look at the processes that

lead to the behavior as well as at the content of the agent’s propositional attitudes. With respect to the

propositional attitudes, these theories focus in particular on the content that represents the sources and

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

processes that lead to the development or alteration of the agent’s behavior-underlying states. According to

a mixed theory of autonomy an agent cannot be autonomous with respect to some bit of behavior if she does

not have (inter alia) non-negative attitudes about the processes leading up to the states that underlie this

behavior. In other words, the agent must approve of the processes. This relation between an agent’s

attitudes and the processes that lead to her behavior plays the same role in the mixed account that the

relation between lower-order and higher-order attitudes plays in the internalist account. That is, it is

meant to ensure that in order to be autonomous an agent must in some sense authorize the forces that move

her. But unlike internalist theories, mixed accounts require that the salient propositional attitudes have as

their content the processes that lead to or underlie the relevant behavior.

11

p. 86

Accounts of manipulation that appeal to an externalist conception of autonomy are di�cult to assess

because manipulation is itself a historical (i.e., external) process, one that is often construed as being

antithetical to autonomy by de�nition. In order to defend my claim that the presence of manipulation is at

least sometimes consistent with autonomous behavior I will have to pursue one of two courses. The �rst is

to argue that externalist theories of autonomy fail and so it does not matter that according to these theories

manipulation and autonomy are inconsistent. This would leave the internalist theory standing and (as

sketched above) autonomy as the internalist construes it is consistent with manipulation. The alternative

approach is to show that manipulation does not always threaten autonomy as understood by externalist

theories. I will pursue the latter strategy for two reasons. The �rst is methodological. I do not want the

plausibility of my account of manipulation to depend on the truth of a controversial theory of autonomy.

Second, I happen to think history does matter when it comes to autonomy. Some of the standard objections

to purely internalist theories are decisive in the absence of any appeal to externalist considerations (i.e.,

historical processes). However, my claim that manipulation is consistent with autonomy does not require

that externalist theories are true but only that, if they are true, it is not clear that they can easily rule out

manipulation as an autonomy-respecting form of in�uence.

12

13

I will take two routes toward supporting the claim that manipulation is consistent with externalist

conceptions of autonomy. The �rst will be to provide a case of manipulation in which, intuitively, no one’s

autonomy is undermined. Next, I will argue in more general terms that the most plausible kind of

externalist theory of autonomy cannot exclude manipulation. Here is a case of manipulation in which

intuitively no one’s autonomy is undermined:

p. 87

Cafeteria: Concerned about skyrocketing obesity rates, the manager of a cafeteria wants his

customers to eat healthful food. One way to do this is by getting his customers to choose salads

instead of french fries. Suppose people tend to choose the items they encounter �rst—that is,

those placed at the front of the food line. Knowing this, the manager places the salads at the front

of the line and places the french fries farther down. Consequently, more people begin to choose the

salads, just as the manager intended. In this case, at least some customers—those who would have

selected the french fries had these been placed at the front of the line—are manipulated into

choosing the healthful items. Yet, intuitively no one’s autonomy is undermined.

14

Cafeteria shows that a person’s autonomy can remain intact despite the presence of manipulation in the

history of the behavior whose autonomy is in question.

There are general arguments to the conclusion that the most plausible theory of autonomy is a mixed theory

and that such theories, like internalist theories and pure externalist theories, render autonomy consistent

with manipulation. With respect to the �rst half of this claim, in order to accommodate some strong

intuitions about autonomy, intuitions regarding the importance of the agent’s attitudes about her own

agential capacities, a defender of an externalist account of autonomy cannot appeal to just those processes

that underlie the agent’s behavior. This is because even if these processes are free from problematic

external interference an agent who is alienated from these processes will lack a critical component of

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

autonomous agency. She will not conceive of herself as an agent acting independently of problematic

interferences.

In the absence of the satisfaction of an attitudinal condition an agent may meet pure externalist conditions

for autonomy and yet falsely believe she is being controlled by autonomy-undermining forces. Or she may

be free of any problematic external interferences and yet lack a coherent set of attitudes—that is, she

may not identify with her lower-order attitudes. It may be a necessary condition for self-governance that an

agent has a conception of herself as self-governing. It is plausible that an agent’s attitudes about her own

agency partly determine the extent to which she actually is an agent, and thus an analysis of autonomous

agency must make some appeal to an agent’s representations of and attitudes about her situation. If this is

right, a defender of an externalist theory of autonomy is pushed towards a mixed theory, a theory that

incorporates some attitudinal condition such as the condition requiring that an agent approve of the

processes that lead up to her behavior.

15

p. 88

Thus far I have tried to motivate externalism about autonomy and I have sketched some of the reasons why

an externalist might be pushed towards a mixed theory. It still remains to be argued that mixed theories

render manipulation consistent with autonomy. Here I will draw from the literature on autonomy and in

particular from work that is critical of internalist theories. As already noted, one of the most powerful

objections to internalist theories is that higher-order and lower-order attitudes can be brought to cohere in

any number of ways, some of which are manipulative. My strategy will be to show that the attitudinal

condition in mixed theories— that is, the condition requiring that an agent have the right sort of attitudes

about the processes leading to her behavior—is vulnerable to the same problem. It will be easier to see this

with an example. Take John Christman’s analysis of autonomy:

(i) A person P is autonomous relative to some desire D if it is the case that P did not resist the

development of D when attending to this process of development, or P would not have resisted that

development had P attended to the process;

(ii) The lack of resistance to the development of D did not take place (or would not have) under the

in�uence of factors that inhibit self-re�ection; and

(iii) The self-re�ection involved in condition (i) is (minimally) rational and involves no self-

deception.16

Whether or not an agent resists the development of some propositional attitude (condition [i] ) is going to be

determined (at least partly) by her other propositional attitudes, so a question arises as to whether the agent

resisted the development of these attitudes. The same question then arises with respect to the attitudes

that determined whether the agent resisted those attitudes. And so on. Condition (ii) may be meant to stop

the regress but it can do so only with respect to methods that inhibit self-re�ection (e.g., brainwashing).

Other methods, such as presenting an agent with a circumscribed set of options, presenting those options in

one order rather than another, or even creating a context in which an agent is more likely to be self-

re�ective are not ruled out by the condition speci�ed in (ii).

p. 89

17

As far as I can tell there is no way for an account of autonomy that incorporates an attitudinal condition to

exclude manipulation. The only way to exclude manipulation is by jettisoning the attitudinal condition and

sticking with a pure externalist view. However, as I have already suggested, I do not think pure externalist

accounts of manipulation work. (And even if they do work qua theories of autonomy, cases like Cafeteria

suggest that such theories may not be able to rule out manipulation). Therefore, the most plausible

competing accounts of autonomy—the internalist account and the mixed externalist account—construe

autonomy in manner that renders it consistent with the presence of manipulation.

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

Manipulation and the Rational Capacities

Perhaps the most popular account of manipulation is the one according to which manipulation essentially

bypasses or subverts the rational capacities of the manipulated agent. I will refer to this account as the

Bypass or Subvert View (BSV) of manipulation. I have argued elsewhere that BSV is false and do not have the

space here to recite these arguments in any detail. Thus, I will recapitulate only the main claims and

then go on to situate them in relation to the examples provided above in earlier sections.

18p. 90

According to BSV, manipulation is a process of interpersonal in�uence that necessarily either fails to engage

the rational capacities of the in�uenced agent (that is, it bypasses them) or else engages these capacities in

some way that undermines their function (that is, it subverts them). BSV captures science-�ctional cases of

manipulation of the sort frequently discussed in the free-will literature—for example, cases of direct

neurophysiological interventions or high-tech brainwashing. It also captures more realistic cases in which a

manipulator stimulates psychological states in the manipulee that interfere with the manipulee’s ability to

re�ect on her situation or to act in light of her better judgment.

BSV can account for many cases of manipulation—probably most of them—and it can explain why

manipulation is often associated with “pulling the strings” of the manipulee, with conceiving of others as

“puppets” to be managed or handled. It can also explain the tendency to associate manipulation with

deception and with the undermining of autonomy. Thus, BSV is quite plausible and has been advanced in

one form or another by a number of philosophers writing on manipulation. Writing in this volume, Allen

Wood suggests that manipulation “in�uences people’s choices in ways that circumvent or subvert their

rational decision-making processes, and that undermine and disrupt the ways of choosing that they

themselves would critically endorse if they considered the matter in a way that is lucid and free of error.”

At one time, BSV was defended by two other contributors to this volume, Marcia Baron and Eric Cave. Baron

concludes her (earlier) essay on manipulativeness with the claim that “manipulativeness...re�ects a failure

to view others as rational beings or an impatience over the nuisance of having to treat them as rational.”

In his (earlier) paper, Eric Cave defends the claim that “successful motive manipulation...operates by

bypassing an agent’s capacities for control over her mental life.” Patricia Greenspan has also defended

BSV, claiming that

19

20

21

p. 91 22

[A]s an interference with rational self-governance, the problem with manipulation is not just that it

fails to respect some notion of personal separateness or boundaries between persons but rather

that it undermines a basic assumption of interpersonal trust in groups that make claims on their

members on grounds of fraternity among reasoning agents.23

In her essay on manipulation and politics, Claudia Mills suggests that manipulators in�uence manipulees by

“o�ering...bad reasons, disguised as good, or faulty arguments, disguised as sound.” Scanlon, too, seems

committed to something like BSV. As noted earlier, Scanlon’s Principle M is meant to rule out one agent’s

leading another agent to form expectations based on intentionally misleading behavior on the part of the

�rst agent. Again, on his view Principle M is a valid moral principle because “[c] onsidering the matter from

the point of view of potential victims of manipulation, there is a strong generic reason to want to be able to

direct one’s e�orts and resources toward aims one has chosen and not to have one’s planning co-opted.”

Insofar as the rational capacities play a central role in helping an agent direct her energies towards aims she

has chosen, manipulation subverts these capacities.

24

25

Despite its plausibility I think BSV is false. Consider the following case:

Election: Jones is running for president. He has never taken the time to educate himself about

matters of public concern and in fact could not care less about politics. Jones is attracted to the

presidency because he intensely craves attention and knows that those who occupy this o�ce are

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

constantly in the spotlight. In order to make himself politically salable Jones hires a stable of

pollsters, speech writers, and acting coaches. He constructs a political platform with the maximal

vote-getting potential, as determined by his polling experts. He also commits to following through

on his campaign promises once elected, since he calculates that this may help him win a second

term and hence to extend his time as the center of attention. Though personally indi�erent to the

ideals re�ected in the policies he publicly defends, Jones—with the help of his speechwriters and

coaches—convincingly advocates for his platform in speeches and debates and garners enough

support to win the election.

p. 92

In Election, Jones behaves manipulatively. Nevertheless, Jones does not bypass or subvert the rational

capacities of the citizens who vote for him. After all, Jones advocates in favor of policies that many citizens

really do support and he commits himself to realizing the values of these citizens. Jones does not deceive

anyone about what he intends to do when elected and those who vote for him do so on the basis of political

preferences upon which (we may suppose) they have rationally settled.

BSV holds that manipulation always involves the bypassing or subversion of the manipulee’s rational

capacities, and yet Election is an example of manipulation in which the rational capacities of the

manipulated agent are neither bypassed nor subverted. Therefore, despite its intuitive appeal and ability

both to account for many cases of manipulation and to subsume other competing theories, BSV is false.

Manipulation and Reasons

I believe the plausibility of the accounts discussed so far—and of BSV in particular—is grounded in our

sense that manipulation does not track reasons in the way some other less problematic forms of

interpersonal in�uence do. That is, I think it is manipulation’s failure to track reasons that renders it

morally suspect. In this section I will provide a rough account of how I think manipulation fails to track

reasons.

As a �rst step, consider Mills’s insightful observation that

A manipulator judges reasons and arguments not by their quality but their e�cacy. A manipulator

is interested in reasons not as logical justi�ers but as causal levers. For the manipulator, reasons

are tools, and a bad reason can work as well as, or better than, a good one.26

I think Mills gets something right here, but also that she overlooks a critical implication of her claim and

thus misses something central about the nature of manipulation, namely that for those who care only about

the causal e�cacy of reasons or arguments—that is, for manipulators—the degree to which a particular

reason or argument provides justi�catory support will be irrelevant, and this will be as true with respect to

good reasons and arguments as it is with respect to bad ones. Mills puts too much emphasis on the provision

of bad reasons and arguments and consequently misses the crucial point, which is that for manipulators the

justi�catory quality of their means of in�uence just does not matter.

p. 93

This point can be illuminated by contrasting the behavior of Jones in Election with the behavior of Smith,

who is running for president in a nearby possible world. Smith and Jones are physically indistinguishable

from one another and they run qualitatively identical campaigns from the point of view of voters. The

di�erence between Smith and Jones is that Smith really does believe in the wisdom of the policies he

proposes and he really does adhere to the values he believes will be realized by these policies. Smith wants to

become president because he believes the policies he defends will be good for his country and its citizens.

Most important, Smith believes the considerations that count in favor of his policies provide citizens with

good (justifying) reasons to support his candidacy, and the reason-supportedness of these policies is what

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

ultimately motivates Smith to seek public o�ce. Put another way, the reasons Smith believes support his

candidacy are the same reasons to which he appeals when he asks voters for their support. Unlike Jones, for

whom the causal e�cacy of his advocacy is the only aspect he considers, if Smith were to realize that one of

the policies he favors is misguided he would make the relevant adjustments to his platform.

The salient di�erence between Smith and Jones—the feature that renders only the latter’s behavior

manipulative—is that the consideration that motivates Jones to run for o�ce (i.e., that it will bring him a

lot of attention) is completely independent of the consideration he invokes in trying to motivate voters to

support him (i.e., that the policies he will implement are the best policies). For Jones the fact that there is

good reason for a voter to see some policy implemented does not play a role in the explanation of why he

publicly seeks to implement that policy—recall that he chooses policies strictly on the basis of their

popularity. Similarly, in Flattery when Carlos compliments his boss Lucinda he is not motivated by his

beliefs regarding the features of Lucinda’s behavior that really do justify a compliment but, instead, by

his desire to get into her good graces. That what he says happens to be supported from Lucinda’s point of

view by good reasons plays no role in the explanation of his behavior.

p. 94

Both Jones and Smith intend to bring about behavior that is, from the perspective of voters, supported by

good reasons. However, only Smith is motivated by the reasons he believes really do support the behavior he

seeks to bring about, while Jones is completely indi�erent to the justi�catory dimension of the reasons to

which he appeals, caring only about their causal e�cacy. Thus, though Jones and Smith both intend to bring

about reason-supported behavior, and though both Jones and Smith appeal to these reasons in asking for

support, only Smith intends to bring about the reason-supported behavior because it is supported by

reasons and only Smith would revise his plans if he were to come to believe that the behavior at which he is

aiming is not, in fact, supported by reasons. The process of in�uence initiated by Smith tracks reasons; the

process initiated by Jones does not.

In Election, the behavior at which Jones aims is in relation to those he seeks to in�uence supported by

reasons, though this does not matter to Jones and plays no role in forming his intentions. I call

manipulation of this type non-paternalistic reasonable manipulation (NPM). In other cases of manipulation,

however, the reason-supportedness of the behavior does play a role in the intentions of the manipulator—

that is, the manipulator aims at reasonable behavior because it is, from the perspective of the manipulee,

reasonable. I call this sort of manipulation paternalistic manipulation (PM). Dementia, in which a concerned

son manipulates his demented mother into admitting herself to the hospital for treatment, is a case of

paternalistic manipulation. PM, like NPM, is reasonable—that is to say, the manipulator believes the

behavior that is the upshot of the manipulation is, from the perspective of the manipulee, supported by

good reasons. The di�erence between PM and NPRM lies in the motivational base of the manipulator.

Paternalistic manipulators are motivated by the reason-supportedness of the behavior, while for non-

paternalistic manipulators like Jones the reason-supportedness of the behavior is merely incidental and

plays no direct role in explaining their intention to in�uence others.

However, despite aiming for behavior that is supported by reasons because it is supported by reasons,

paternalistic manipulators still make use of means of in�uence that fail to track reasons. For example,

Nathaniel encourages his mother, Mildred, to mistake him for his father. Nathaniel knows this belief does

not justify his mother’s going to the hospital for treatment. He makes use of his mother’s belief solely

because of the causal role he knows it can play and not because the content of her belief provides

justi�catory support for her action. Similarly, many of the “paternalistic libertarian” forms of in�uence

defended by Thaler and Sunstein exploit cognitive biases such as the framing e�ect, anchoring, the default

bias, and so on, which are psychological processes that do not track reasons. Paternalistic manipulators

typically choose manipulation only when reason-tracking forms of in�uence are unlikely to work and when

they believe the reasons that support the behavior they seek to bring about are su�ciently weighty to justify

non-reason-tracking means of in�uence.

p. 95

27

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

Thus far I have discussed two types of manipulation, both of which involve manipulees being in�uenced in

ways that lead to their behaving in reason-supported ways. But obviously there are many cases of

manipulation that do not �t into either category because the manipulator is aiming at behavior he believes

to be unsupported by reasons. Unreasonable manipulation (UM) is a form of in�uence in which an in�uencer

aims to get the person being in�uenced to behave in ways the in�uencer believes to be unsupported by

reasons, viewing the matter from the point of view of the manipulee.28

The sort of example that motivates the Harm-Based View of manipulation are also (typically) cases of UM.

As I suggested when discussing the Harm-Based View, if an agent is su�ciently rational it will be di�cult

or impossible to get her to behave in ways she believes to be unsupported by good reasons. In such cases, a

determined in�uencer who also believes the behavior she is seeking is unsupported by good reasons cannot

appeal to what she thinks are good reasons. She might present bad reasons as good ones or she may

bypass the manipulee’s rational capacities altogether, opting instead to stimulate a compulsion or to appeal

to a misdirected or out-of-proportion emotion. UM fails to track reasons because it is not part of the

manipulator’s intention that the manipulee behave reasonably—on the contrary, here the manipulator

consciously aims at behavior she believes is unsupported by reasons.

p. 96

I have argued—in an admittedly sketchy way—that interpersonal manipulation, whether it is UM, PM, or

NPM, fails to track reasons. Sometimes a manipulator will try to get others to behave in ways the

manipulator believes are unsupported by reasons. Sometimes a manipulator will aim for behavior that

happens to be supported by reasons, though the reason-supportedness of the behavior does not provide an

independent motivational base for the manipulator. And �nally, there are times when the reason-

supportedness of the behavior does provide a source of motivation for the manipulator, though here the

manipulator makes use of means of in�uence that do not track the supporting reasons.

On this view, the most salient elements of an interpersonal encounter in which one agent seeks to in�uence

another are the motivations of the in�uencer, the particular means of in�uence chosen by the in�uencer,

and the propositional attitudes and other mental states of the agent being in�uenced. In an ideal type of

interpersonal in�uence, like rational persuasion with the right intentions, everything links up nicely: the

motivations of the in�uencer are grounded in the reasons she believes really do support the behavior she

seeks to bring about, the means of in�uence (e.g., sound argument) reliably “aim at” or “link up with”

these reasons, and the mental states of the person being in�uenced also refer to the reasons that support

the behavior. In cases of manipulation, there are breakdowns in these relations, breakdowns that can occur

in more than one place. The location of the breakdown will determine whether the manipulation is

reasonable or unreasonable, paternalistic or non-paternalistic.

Though I cannot argue for this claim here, I believe manipulation’s failure to track reasons is what makes it

morally impermissible (when it is). I believe this is true even when manipulation does not involve

deception, harm, the bypassing or subversion of the rational capacities, or the violation of the manipulated

agent’s autonomy. Processes of in�uence that fail to track reasons can leave those who are in�uenced by

them detached from an important aspect of reality, namely the considerations that ought to govern their

behavior—their reasons. In�uencing reasons-responsive agents via such processes displays a lack of

respect for them as agents, and this is pro tanto wrong.

p. 97 29

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

Notes

Conclusion

I have argued that although it often does so, manipulation does not necessarily involve common wrongs like

deception, harm, the bypassing or subversion of the rational capacities, or the violation of the manipulated

agent’s autonomy. Some of the accounts I have criticized—for example, the Harm-Based View—were not

very plausible to begin with. Others are much more attractive. The Bypass or Subvert View in particular is

quite plausible, at least initially, and a number of philosophers have defended it. However, despite its

covering probably the majority of cases of manipulation, BSV cannot capture all of them. Consequently, we

are need of an alternative account of interpersonal manipulation.

I then went on to sketch an alternative theory. According to this account manipulation is a process of

interpersonal in�uence that deliberately fails to track reasons. Sometimes manipulators intend their

manipulees to behave in ways they (the manipulators) do not believe to be supported by reasons. Sometimes

manipulators aim at behavior they do believe to be supported by reasons, but here either they remain

unmotivated by these reasons or they make use of means of in�uence that do not reliably track those

reasons. Finally, I suggested that manipulation’s failure to track reasons can help explain why manipulation

is wrong (when it is wrong). Clearly much more needs to be said, both about what manipulation is and about

what makes it wrong. For now, I hope to have accomplished two relatively modest things. First, I hope to

have shown why some accounts of manipulation that initially might appear attractive—those that appeal to

phenomena such as deception, harm, autonomy, and the rational capacities—do not work. Second, I have

provided a sketch of what I believe to be a better theory, one that begins to provide a general explanation of

the nature and ethical signi�cance of interpersonal manipulation.

1 Sarah Buss, “Valuing Autonomy and Respecting Persons: Manipulation, Seduction, and the Basis of Moral Constraints,”

Ethics 115 (January 2005): 226 10.1086/426304 . 2 These expectations correlate roughly to the four maxims constituting Griceʼs Cooperative Principle (quantity, quality,

relation, and manner). Grice was attempting to provide a theory of meaning in formulating the Cooperative Principle and in examining various failures to abide by the Principle. I do not mean to endorse Griceʼs semantic theory. I appeal to Griceʼs categories here because they are helpful in articulating the kind of expectations that are generated in a wide range of social interactions. For Griceʼs discussion of the Cooperative Principle, see H. P. Grice, “Logic and Conversation,” in Studies in the Way of Words (Cambridge, MA: Harvard University Press, 1989), 22–40.

3 Thomas Scanlon, What We Owe to Each Other (Cambridge, MA: Belknap,, 1998), 298. 4 Scanlon, What We Owe to Each Other. I think it is plausible that Principle M is indeed a valid moral principle. However, the

principle is formulated in such a way as to preclude more than one kind of morally questionable behavior, and thus it is not clear that it best accounts for the wrongness of manipulation rather than some other kind of wrong. First, as Scanlon points out, agent A makes it impossible for B to “direct [his or her] e�orts and resources toward aims [B] has chosen.” Second, A has intentionally sought to gain an advantage at Bʼs expense, as we are told B will su�er significant loss. Third, A has deceived B about Aʼs intentions, the content of which intentions form the basis of Bʼs decision to behave as A wishes. None of these three things forms an essential component of the others—they are conceptually independent. One might commit one of these putative wrongs without committing the others.

5 Buss, “Valuing Autonomy and Respecting Persons,” 226. 6 Scanlon, What We Owe to Each Other 299–322. 7 Carol Rovane, The Bounds of Agency: An Essay in Revisionary Metaphysics (Princeton, NJ: Princeton University Press, 1998),

78. 8 Richard Thaler and Cass Sunstein, Nudge: Improving Decisions About Health, Wealth, and Happiness (New York: Penguin,

2008); Cass R. Sunstein and and Richard Thaler, “Libertarian Paternalism Is Not an Oxymoron,” University of Chicago Law

Review 70 (2003): 1159–1202 10.2307/1600573 . 9 Sunstein and Thaler, “Libertarian Paternalism,” 1184. 10 See, for example, Harry Frankfurt, “Freedom of the Will and the Concept of a Person,” Journal of Philosophy 68, no. 1

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024

(January 1971): 5–20 10.2307/2024717 . 11 I understand ʻapproveʼ rather weakly as a kind of (actual or perhaps even hypothetical) pro-attitude. 12 Here I have in mind certain counterexamples to internalism. Mele provides some powerful ones in Alfred Mele,

Autonomous Agents: From Self Control to Autonomy (Oxford University Press, 2001) 10.1093/0195150430.001.0001 . 13 I thank George Sher for pointing out that my arguments regarding manipulation and autonomy can remain neutral on the

question of which sort of theory of autonomy—internalist or externalist—is the one we ought ultimately to adopt. 14 Sunstein and Thaler, “Libertarian Paternalism,” 1164. They discuss the same example in Thaler and Sunstein, Nudge 1–4. 15 That is to say, the history of how she came to be in the states she is in and to have the attitudes she has may include no

external interferences that obviously threaten autonomy (e.g., brainwashing). 16 John Christman, “Autonomy and Personal History,” Canadian Journal of Philosophy 21, no. 1 (1991): 11. 17 I ignore the hypothetical versions of Christmanʼs analysis. First, I am not yet sure how to interpret them. Second, unless

the hypothetical consent is the consent of a idealized agent, hypothetical consent seems to reduce to some set of facts about the actual agent that are quite independent of issues of consent. In short, I am generally skeptical about the ability of hypothetical consent to render an agent autonomous.

18 Moti Gorin, “Do Manipulators Always Threaten Rationality?,” American Philosophical Quarterly 51, no. 1 (January 2014). 19 See Marcia Baron, “Manipulativeness,” Proceedings and Addresses of the American Philosophical Association 77, no. 2

(November 2003): 50 10.2307/3219740 ; Tom Beauchamp and J. Childress, Principle of Biomedical Ethics, 6th ed. (Oxford: Oxford University Press, 2008): 133–34; Jennifer Blumenthal-Barby and Hadley Burroughs, “Seeking Better Health Care Outcomes: The Ethics of Using the Nudge,” American Journal of Bioethics 12, no. 2 (2012): 5 10.1080/15265161.2011.634481 ; Eric Cave, “Whatʼs Wrong with Motive Manipulation?” Ethical Theory and Moral

Practice 10, no. 2 (2007): 138 10.1007/s10677-006-9052-4 ; Patricia Greenspan, “The Problem with Manipulation,” American Philosophical Quarterly 40, no. 2 (Apr. 2003): 164; Claudia Mills, “Politics and Manipulation,” Social Theory and

Practice 21, no. 1 (Spring 1995):100 10.5840/soctheorpract199521120 ; Lawrence Stern, “Freedom, Blame, and Moral Community,” Journal of Philosophy 71, no. 3 (February 1974): 74 10.2307/2024987 .

20 Allen Wood, “Coercion, Manipulation, Exploitation,” this volume,xx–xx. 21 Baron, “Manipulativeness,” 50. 22 Cave, “Whatʼs Wrong with Motive Manipulation?” 133. 23 Greenspan, “The Problem with Manipulation,” 164 (emphasis mine). 24 Mills, “Politics and Manipulation,” 100. 25 Scanlon, What We Owe to Each Other, 298. 26 Mills, “Politics and Manipulation,” 100–101. 27 Sunstein and Thaler, “Libertarian Paternalism,” 1159–202. Cognitive biases do not track reasons because the explanation

for why an agent chose as she did—suppose, for example, an agent is much more likely to choose option A than option B when option A is presented as the default option, even when option B is superior—makes no reference to reasons that could be invoked to justify that choice.

28 The account I am sketching here is similar to Robert Noggleʼs account of manipulative action. According to Noggle, “manipulative action is the attempt to get someoneʼs beliefs, desires, or emotions to violate...norms, to fall short of...ideals” where the norms in question are the norms of belief, desire, and emotion, and the ideals are those of the influencing agent. I disagree with Noggle on several points, though I do not have the space here adequately to engage with his view. My main worry with Noggleʼs construal of manipulation is that it is not clear that it can easily account for paternalistic manipulation. Robert Noggle, “Manipulative Actions: A Conceptual and Moral Analysis,” American Philosophical Quarterly 33, no. 1 (January 1996): 44.

29 I discuss the ethics of manipulation in an unpublished manuscript.

D ow

nloaded from https://academ

ic.oup.com /book/4870/chapter/147239683 by Florida State U

niversity user on 02 January 2024