Professional Documents
Culture Documents
Sarah Buss
* So many people have offered me helpful comments on this article, and I have done
my best to respond to them all. As always, I am grateful for several consulting sessions
with Maggie Little. I would also like to thank Jennifer Hawkins, Elijah Millgram, Martha
Nussbaum, Angela Smith, Gopal Sreenivassan, David Velleman, Carol Voeller, Michelle
Kosch, and especially Marcia Baron. I have benefited greatly from the comments of eight
anonymous readers and from the opportunity to discuss the article with audiences at
Vanderbilt University, Stanford University, the University of Calgary, the University of Iowa,
and the University of Illinois, and with the members of the short-lived but lovely-while-it-
lasted PHILAMORE.
1. Thus, Thomas Nagel summarizes Rawls’s Kantian thesis as follows: “The funda-
mental attitude toward persons on which justice as fairness depends is a respect for their
autonomy” (“Rawls on Justice,” in Other Minds [New York: Oxford University Press, 1995],
pp. 121–36, p. 125).
195
196 Ethics January 2005
2. Of course, as Kant took pains to point out, whether we are dealing in concepts or
conceptions, the conclusions will be conditional, unless we can also establish that we are,
in fact, autonomous agents: they will be conclusions about what we ought to do, if we
have the capacity to govern ourselves when we act. As should be clear, for the purposes
of this article, I am taking it for granted that on any adequate account of autonomy, most
of us do, in fact, have this capacity, most of the time.
198 Ethics January 2005
3. The distinction I draw here is, I think, quite similar to the distinction that John
Christman has in mind when he notes that “accounts which explicate the relation between
autonomy and principles of justice and rights should be separated into two different
categories. On the one hand, writers like Rawls and Gewirth—both heavily influenced by
Kant—see autonomy as one principal property of persons that determines their ability to
derive the principles of morality and justice. . . . This can be contrasted with the view
that autonomy is a valuable character trait of individuals which, from the point of view
of the moral theorist, should be afforded respect and/or protection. . . . Autonomy, on
[this second] view . . . , provides the basis for the right to be treated as a free and equal
moral person, a fundamental human right. . . . The characteristics of persons that warrant
the root idea that they are owed basic respect (no one can be used as a resource for
another) is their capacity for rational agency and the ability to formulate a plan of life.
. . . For both Richards and Nozick, autonomy is seen as the characteristic of persons
whose value (from the point of view of the theorist) is only protected properly when a
certain set of human or natural rights is attributed to the agent and respected by others.”
According to Christman, the first approach “amounts to construing autonomy as a kind
of moral neutrality by which agents can be said to construct or derive moral principles
from their point of view” (John Christman, “Constructing the Inner Citadel: Recent Work
on Autonomy,” Ethics 99 [1988]: 109–24, p. 119). To my mind, however, this construal of
autonomy reflects something deeper: a conception of autonomy as government by reason
and of reason as “neutral” with respect to any but its own intrinsic ends.
4. One element of the controversy concerns the relationship between the Kantian
“ideal” of self-government and the sort of self-determination that is characteristic of im-
moral agents. Whatever the details, the rough idea is that insofar as our maxims are
incompatible with the Categorical Imperative, we are violating the law of our own will
and are, to this extent, governed by a law that is “external,” or “alien,” to our will. Of
course, even when we set this issue to one side, there is the additional difficulty of ex-
plaining just how the Categorical Imperative rules out manipulation and deceit.
Buss Valuing Autonomy and Respecting Persons 199
Thomas Hill explains, this means that the capacity for self-government
“is attributed only to those presumed to have certain basic normative
capacities and dispositions. These include the ability to reflect on one’s
desires and circumstances, to set ends for oneself, to form coherent
plans.”6
To this list Hill adds the disposition “to be willing to reciprocate
with others in endorsing principles that respect each person as a po-
tential source of legitimate ends.”7 He appears to take this disposition
to be an intrinsically moral capacity. But this is precisely what I want to
deny. With Hill and other Kantians, I begin with the assumption that
“when we acknowledge, as we must, the . . . value of rational agency
in others, what we grant as important per se is not their happiness or
particular projects but their existence, development, and opportunity
to live as rational agents.”8 Having conceded this much, however, I want
to challenge the further assumption that in order to “grant as important
per se” other persons’ “existence, development, and opportunity to live
as rational agents,” one must adopt a moral point of view. I am prepared
to grant that we cannot acknowledge the value of other human beings
without acknowledging the value of their developing and exercising
their capacity for self-determination. But as far as I can tell, there are
nonmoral, and even immoral, ways of interpreting this evaluative
commitment.
Consider Kierkegaard’s aesthetes. They govern themselves in ac-
cordance with their own conceptions of what is, not morally good, but
aesthetically good, that is, interesting. Like the paradigmatic moral
agents, however, they regard the capacity to govern oneself as an end
in itself; and so they regard everyone who has this capacity as a “legit-
imate source of ends.” Everything a true aesthete does is conditional
upon his acknowledging the value of the capacity for self-government.
He is dedicated to the proposition that this capacity—the capacity to
evaluate the merely given in terms of one’s own criteria, to make some-
thing of it in one’s own way, to respond to it according to one’s own
son”(Douglas N. Husak, “Paternalism and Autonomy,” Philosophy & Public Affairs 10 [1980]:
27–46, p. 28); (3) “To the extent that individuals survey their options guided by their self-
scrutinized feelings, values, goals, and the like, and then marshall the determination to
follow their own counsel, they live autonomously” (Diana Meyers, “Personal Autonomy
and the Paradox of Feminine Socialization,” Journal of Philosophy 84 [1987]: 619–28, p.
627). For further examples, see n. 25.
6. Thomas Hill, “Respect for Humanity,” in The Tanner Lectures on Human Values, vol.
18, ed. Grethe B. Petersen (Salt Lake City: University of Utah Press, 1997), pp. 1–76, p.
37.
7. Ibid.
8. Thomas Hill, “Kant’s Theory of Practical Reason,” in Dignity and Practical Reason
in Kant’s Moral Philosophy (Ithaca, N.Y.: Cornell University Press, 1992), pp. 123–46, p. 145.
Buss Valuing Autonomy and Respecting Persons 201
entries and by interacting with other agents in ways that “bring it out.”
By means of virtuosic feats of self-reflection and self-control, he seduces
Cordelia into valuing this capacity in herself as well as in him.
And then he leaves her. Without regret, without any concern for
her feelings, he abruptly cuts off all communication with her because,
he believes, he has done all that he can to honor the unconditional
value of her autonomy. He has exhausted the opportunities for enjoying
and developing her ability to think for herself, to eschew the restrictions
of external authority, to contemplate alternate possibilities, and in other
ways to exercise her capacity for self-government. He always knew that
this day would come. Cordelia has already broken their engagement on
the grounds that it is an unacceptable restriction of her freedom; and
he takes the same attitude toward all long-term commitments, including
the long-term commitment of marriage. More generally, he is opposed
to relationships of mutual dependence, which, in his view, narrow the
scope of a rational agent’s powers of self-legislation by establishing en-
trenched habits of thought and evaluation whose influence is not under
the agent’s control. And so, he believes, he must terminate the rela-
tionship if he is to remain true to his values.12
Since Cordelia is a less thorough aesthete than he is, his decision
to abandon her—and especially his abrupt manner of severing all ties
with her just when she has broken all other ties in order to become his
lover—makes her miserable. But it does not compromise her newly
developed autonomy.13 And it is her autonomy—not her happiness or
her “particular projects”—that he values as an end in itself. He values
her as an end in herself insofar as, and only insofar as, she is an au-
tonomous agent.
One might plausibly object to my characterization of the Seducer’s
motives. After all, he seems to contradict himself from one diary entry
to the next. And Kierkegaard wants us to conclude that he is, ultimately,
self-deceiving—that he is not as free as he thinks he is, precisely because
he refuses to acknowledge the constraints imposed by external reality
and so is constantly broadsided by these constraints. If, however, the
motivation I have attributed to him is different from the one that really
12. Jennifer Hawkins has called my attention to similarities between the case of Jo-
hannes and Cordelia and the case of the two main characters in the movie Dangerous
Liaisons (and the book on which the movie is based). These two people, she suggests,
“are aesthetes of sorts.” “They enjoy manipulating others. But they admire each other so
much precisely because they are the two most autonomous people they know. . . . But
precisely because this is what they admire about one another, they can’t really come
together as lovers. . . . If one of the pair started to trust the other, then the one in control
would suddenly see the other as having sacrificed autonomy, and so would suddenly despise
that one as weak” (personal correspondence).
13. But see n. 36.
Buss Valuing Autonomy and Respecting Persons 203
16. Søren Kierkegaard, “Equilibrium,” in Either/Or, vol. 2, trans. Walter Lowrie (Prince-
ton, N.J.: Princeton University Press, 1959), pp. 159–338, p. 164.
17. Kierkegaard, Diary of the Seducer, p. 304.
18. Harry Frankfurt, “On the Necessity of Ideals,” in Necessity, Volition, and Love (Cam-
bridge: Cambridge University Press, 1999), pp. 108–17, pp. 114–15.
Buss Valuing Autonomy and Respecting Persons 207
under certain specified circumstances. Finally, I will point out that there
is good reason to think that autonomous agents would countenance
being manipulated or deceived in a wide range of cases. These argu-
ments together suggest that when we are obligated to refrain from ma-
nipulating or deceiving one another, this has relatively little to do with
the value of autonomy. I will bring my case to a close by calling attention
to some valuable conditions that do seem to be relevant to our moral
assessments—conditions whose independent moral significance is too
often obscured by appeals to the moral value of autonomy.
The Effects of Manipulation and Deceit upon the Autonomy of the One
Manipulated and Deceived
What makes Johannes’s treatment of Cordelia so despicable? The ob-
vious answer seems to be that he manipulates and deceives her. Without
ever telling her lies, he induces her to fall in love with him (and with
the life of the imagination) by giving her various “side-glances,” by talk-
ing with her aunt in a way designed to pique her curiosity, by pretending
not to notice her, by leaving a book of poetry open to an especially
moving passage, by speaking in a disparaging way of conventions, by
speaking passionately about freedom, and so forth. He “works on” her
emotions without her knowing what is going on; and he presents her
with a carefully controlled, and seductively partial, image of what he is
really like.
What, exactly, is wrong with this? Why is manipulating someone
incompatible with treating her with moral respect?19 One popular an-
19. It is perhaps worth stressing that, like all those who purport to make more than
tautologous claims in condemning manipulation, I am using the term ‘manipulation’ in
a purely descriptive, nonevaluative sense. My claims thus do not apply to methods of
influence specifically defined as methods that no rational agent could reasonably endorse.
(See, e.g., Thomas Hill’s characterization: “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” [Thomas Hill, “Autonomy and Benevolent
Lies,” Journal of Value Inquiry 18 (1984): 251–67, p. 251].) Clearly, if ‘manipulation’ is
defined as a method of influence that rational people would reject, it is not a defensible
form of influence. The interesting question, however, is: what sort of influence would
rational people reject, and why? And, in particular, would they necessarily reject the sort
of influence that bypasses their rationality in order to get them to see their reasons for
action in a way they might otherwise not? Christman contrasts “manipulative” processes
of preference formation with “‘normal’ processes of self-development” (John Christman,
“Autonomy and Personal History,” Canadian Journal of Philosophy 21 [1991]: 1–24, p. 10).
If, however, the distinguishing mark of manipulative “processes” is that they influence
preferences (and beliefs) nonrationally, then, as I suggest, there is hardly anything more
normal to be found in human interactions. (Note that similar remarks apply to my use
of the term ‘deceit’. When Herman claims that “what counts as deceit varies with the
persons involved, the social conditions of expectations, particular practices, and so on”
Buss Valuing Autonomy and Respecting Persons 209
[“Agency, Attachment, and Difference,” in The Practice of Moral Judgment (Cambridge, Mass.:
Harvard University Press, 1993), pp. 184–207, p. 204], she is conceiving of deceit as having
an essentially negative moral valence. In contrast, when I use the term, I mean to refer
to the intentional misleading of one person by another; and my two [distinct] points
about this way of relating to someone are [1] that it is a far more common feature of
benign human interactions than many moral philosophers seem to think, and [2] when
it is morally wrong to treat someone this way, this has less to do with autonomy than many
moral philosophers seem to think.)
20. As Paul Hoffman puts it, “Seduction is another way in which our will is overpow-
ered. When we are seduced, what we propose to do is not really up to us, it is up to our
seducer” (Paul Hoffman, “Freedom and Strength of Will: Descartes and Albritton,” Phil-
osophical Studies 77 [1995]: 241–60, p. 54).
21. Onora O’Neill, “Between Consenting Adults,” in Constructions of Reason: Kant’s
Practical Philosophy (Cambridge: Cambridge University Press, 1989), pp. 105–25, p. 106.
22. Barbara Herman, “Leaving Deontology Behind,” in The Practice of Moral Judgment,
pp. 208–40, p. 228.
23. Christine Korsgaard, “The Right to Lie: On Dealing with Evil,” Philosophy & Public
Affairs 15 (1986): 325–49, p. 334.
210 Ethics January 2005
his beloved from “choosing for herself which transaction she will engage
in and which ends she will promote.” Of course, Johannes “pulls the
levers” that determine (or at least heavily influence) Cordelia’s thinking
about how to act—and how, in particular, to respond to him; he treats
her “reason as a mediate cause.”24 But this does not mean that he un-
dermines her capacity for reasoning—or even that he prevents her from
exercising this capacity. To the contrary, he takes pains not to do this—
as he must, if he is to avoid forcing her to do things “against her will.”
Those who appeal to the moral impermissibility of “pulling some-
one’s levers,” “pushing her buttons,” treating her as an “instrument” or
“tool,” recognize, of course, that manipulating someone’s will is not the
same thing as compelling her to do something she wills not to do. (They
recognize that there is a difference between rape and seduction.) None-
theless, they claim that manipulation is morally wrong because the ma-
nipulated agent resembles the compelled agent in being deliberately
prevented from exercising the control over her action that is essential
to its being truly her own. Manipulating someone may not prevent her
from reasoning about what to do; but, they insist, it does prevent her
from acting for her own reasons, and this is ultimately why it is wrong.25
24. Ibid.
25. This is what lies behind the view many philosophers share with Bernard Berofsky,
namely, that someone’s “autonomy is respected insofar as his desiring nature as given is
accommodated and the method of influence is restricted to the techniques of rational
persuasion” (Bernard Berofsky, “Autonomy,” in How Many Questions? Essays in Honor of
Sidney Morgenbesser, ed. Leigh S. Cauman, Isaac Levi, Charles Parsons, and Robert Schwartz
[Indianapolis, Ind.: Hackett, 1983], pp. 301–20, p. 311). Though Berofsky does not say
“only insofar as,” he seems to be committed to this further qualification. Christman spells
out this commitment more fully: “For a person to be autonomous, it must be true that
during the processes where she might have resisted these developing desires, she wasn’t
also under the influence of manipulating factors that inhibit the person’s ability to reflect
on her desires and those processes that helped form them. These will be factors whose
presence undermines an agent’s rationality and self-awareness, by virtue of which she
could critically appraise the development of desires and values she finds herself with. [New
paragraph.] The essential aspect of such influences, then, will be whether or not they
crucially affect a person’s reflective capabilities—those cognitive capacities we discussed
above: minimal rationality and self-awareness (non-self-deception). These capacities enable
the agent to appraise and revise certain aspects of herself and the factors which affect her
desires and beliefs. Such things as hypnosis, some drugs, certain educational techniques,
and the like, have the effect of rendering the agent less able to evaluate, from her own
point of view, the processes by which she has come to develop a certain desire or value.
Her vision is clouded, as it were” (Christman, “Autonomy and Personal History,” p. 19).
One of my points in arguing that many cases of nonrational influence are morally benign
is that the critics of manipulation owe us a much more careful account of what it is to
“crucially affect” a person’s “reflective capabilities,” so that she is less able to evaluate
“from her own point of view.” Unless an adult human being is very severely retarded or
demented, she has some ability to “appraise and revise certain aspects of herself and the
factors which affect her desires and beliefs.” What more, exactly, is necessary to qualify
Buss Valuing Autonomy and Respecting Persons 211
her as a competent rational agent? (For a more extensive critique of the attempt to account
for autonomous agency in terms of the ideals of rational agency, see Sarah Buss, “Auton-
omy: Self-Expression in the Passive Mode,” unpublished manuscript. There I also offer
my own account of the sense in which the evaluations of an autonomous agent are “from
her own point of view.”)
26. Barbara Herman, “Murder and Mayhem,” in The Practice of Moral Judgment, pp.
113–31, p. 126.
27. Herman, “Leaving Deontology Behind,” pp. 228, 229, 230.
212 Ethics January 2005
28. Many arguments against incompatibilism make the point that a person can be in
control of her behavior, even if how she exercises this control is causally determined by
some state of affairs, or sequence of events, over which she never had any control. Harry
Frankfurt is among those philosophers who argue that the point does not depend on the
assumption that no rational agent played (intentionally or unintentionally) a causally
determining role in the determining causal chain. Just because one person is morally, as
well as causally, responsible for the fact that a second person makes a certain choice, it
does not follow, Frankfurt says, that the second person is not also responsible. (See Harry
Frankfurt, “Freedom of the Will and the Concept of a Person,” in The Importance of What
We Care About [Cambridge: Cambridge University Press, 1988], pp. 11–25, p. 25, n. 10.)
Buss Valuing Autonomy and Respecting Persons 213
ever they take their actions to be adequately justified, then deceit and
manipulation cannot prevent someone from being a truly autonomous,
truly rational agent.
Deception does not (and cannot) “place the grounds of choice-
worthiness of the victim’s reasons in the deceiver’s will” precisely be-
cause, as Herman herself notes, the victim cannot share the deceiver’s
own point of view on his action. She must act for her own reasons if
she is to act at all; and so she must act for her own reasons if she is to
do what he wants her to do. His success depends not on his making
her take his will as justifying a given choice, but on his acting in such
a way that what he does (and what he wills to do) becomes one of the
factors that causes her to believe—for reasons she herself can defend
“all the way down”—that the choice is, in fact, justified. The fact that
she would have acted differently if she had known more (about his
intentions, in particular) is perfectly compatible with the fact that she
governed herself in light of what she did know.29 If this were not so—
if “transforming reasons into causes” violated the “separateness” of ra-
tional agents—then one would have to be a hermit to preserve one’s
integrity. The salesclerk who wants her commission could, with one false
flattering remark, prevent a shopper from making up her own mind to
buy the unflattering dress. The student who turns in a plagiarized paper
could, just by fooling me into thinking she had written it herself, prevent
me from having my own reasons for giving her an A. The man in Donald
Davidson’s well-known example who plants flowers in order to entice
his neighbor to stop and look could thereby prevent her from making
up her own mind about whether to stop.30
The point, again, is a metaphysical one. I am not here defending
plagiarism, or flattery, or any other favorite example of immorality. I
am not even arguing that deceit and manipulation are sometimes mor-
ally permissible—though I will get to that later. My point is simply that
if such behavior is wrong, this is not because it prevents people from
making up their own minds, determining their own intentions, thinking
for themselves about what is worth doing, and, on this basis, deciding
to do it. If ignorance does not generally have this effect, then there is
ignorance and (2) the judgment that she did not govern her own actions. (As I indicate
later [when I discuss the value of being in touch with reality], confusion is also the result
of failing to distinguish self-government from other desirable conditions.)
36. See Sarah Buss, “Autonomy Reconsidered,” in Midwest Studies in Philosophy 9 (1994):
95–121, and esp. “Autonomy: Self-Expression in the Passive Mode.” The basic idea is that
someone fails to exercise her capacity for rational choice autonomously when she exercises
it in a way that does not express who she truly is, and this happens when the way she
exercises it—the things she takes into consideration in deciding what to do and the relative
“weight” she attributes to these considerations—is decisively influenced by various con-
ditions that are “external”/“alien” to who she truly is, by virtue of being paradigm con-
ditions of human malfunctioning. On this account, autonomy is not merely a component
of a good human life; other components—like the absence of pain, fear, fatigue, depres-
sion—are partly definitive of what autonomy is. This means, of course, that even though
deception and manipulation are not intrinsically autonomy-undermining, they are none-
theless capable of undermining a person’s autonomy by causing her to be afflicted in the
ways mentioned here. It may well be that Cordelia satisfies this condition when Johannes
leaves her. The pain and confusion she feels at being abandoned may be great enough
to prevent her from governing herself. So if, as it is plausible to think, her distress is due,
in part, to the fact that she has been manipulated and deceived, then this manipulation
and deception are partly responsible for her inability to govern herself.
37. Korsgaard, “The Right to Lie,” p. 337.
38. In “Agency, Attachment, and Difference,” Herman herself notes that “if one has
Kantian autonomy as a rational being (able to act on self-given principles), Kantian au-
tonomy may be a necessary but not a sufficient condition for ‘real’ moral autonomy.”
“Ordinarily,” she continues, “we have reason to think that the autonomous person is not
merely one who can act on principles but is, rather, the person whose situation or up-
bringing yields not only a character capable of practically effective critical reflection but
also a character moved by desires and interests that are in some important sense her own”
(pp. 202–3). This leaves us with the task of elucidating the relevant “important sense.”
Herman claims that in this sense “desires and interests that are [either] the result of
216 Ethics January 2005
coercion [or] the products of institutionalized oppression” are not the agent’s own, that
such determining causes—and deceit also—prevent “effective agency” through which “au-
tonomy is expressed (made real)” (p. 205). I do not believe, however, that she tells us
enough about this more robust conception of autonomy to enable us to determine what
it amounts to and in what way it is morally significant.
39. Again, it may be worth reminding the reader of how slippery the terms are—and
how tempting it is to equivocate. There is clearly a sense in which by presenting a person
with reasons for seeing things a certain way, one is treating her as a certain kind of
“instrument”: the kind that responds to reasons.
40. I want to thank Carol Voeller for patiently talking me through the Kantian position
I address in this paragraph. Both she and David Velleman have helped me to see that the
issues touched on here are crucial to evaluating Kant’s understanding of the moral sig-
nificance of autonomy.
41. There are, of course, exceptions. In Survival in Auschwitz, Primo Levi describes a
prisoner named Henri who manipulates the guards by expressing pity for them. It is not
unreasonable to assume that Henri would have reduced these brutal human beings to
mere things if this had been possible, just as he (and many others) would surely have
killed them. (See Primo Levi, Survival in Auschwitz, trans. Guilio Einaudi [New York: Simon
& Schuster, 1996], pp. 98–100.)
Buss Valuing Autonomy and Respecting Persons 217
victims to do their bidding; but this does not mean that they would be
willing to take the steps necessary to disable their victims’ rational agency
in the relevant ways. A willingness to reduce another person to a non-
rational agent—even if only to the point where this would serve one’s
particular purpose—is not an essential component of a willingness to
manipulate and deceive.
I conclude that if manipulating and deceiving someone are gen-
erally wrong, this cannot be because these modes of interacting with
other rational agents either deprive these people of their autonomy
or presuppose a willingness to do so. No independently credible con-
ception of autonomous agency can carry the weight of our belief that
it is wrong to manipulate and deceive people in a wide range of
circumstances.
what transactions they will engage in and what ends they will promote.
This is what it means to treat persons as ends in themselves.”43
On this interpretation of the moral requirement to respect other
people’s capacity for rational choice, Cordelia’s morally relevant in-
ability to consent to Johannes’s intentions is simply the conceptual im-
possibility of her doing so. Deceit and manipulation are always morally
wrong, on this view, because manipulating and/or deceiving someone
is incompatible with her being aware that she is being treated this way.44
This view is appealing in its simplicity. Nonetheless, I believe that
it rests on an exaggerated sense of the desirability of being in control
of one’s own choices. Once we remind ourselves of the wide range of
circumstances under which we are not the least bit threatened by forces
that limit our control over ourselves, we can appreciate, not that ma-
nipulation and deceit are generally acceptable forms of human inter-
action, but that when they are impermissible this has very little to do
with the intrinsic value of the power to think for oneself, to make up
one’s own mind, to act for one’s own reasons.
There are at least two reasons for rejecting the claim that we fail
to respect someone whenever our intentions concerning her are con-
ceptually impossible for her to endorse at the time. First, no one can
consent to any influence on her deliberations without, at the time she
consents, considering (and in at least a minimal sense, assessing) this
43. Christine Korsgaard, “Two Arguments against Lying,” in Creating the Kingdom of
Ends, pp. 335–62, p. 347.
44. Ibid.: “Even the most rational and independent cannot genuinely consent to
proposals about which they are deceived or with which they are compelled to comply.
Even if a proposal would have been welcomed, and coercion or deception is otiose, its
enforcement or surreptitious imposition precludes consent.” See also O’Neill, “Between
Consenting Adults,” p. 111: “The other person cannot assent to my action because he is
not in a position to. This is because he is deceived. By the nature of the case, he doesn’t
know how I am acting, and you can’t assent to a transaction you do not know is occurring.
In the same way, he cannot ‘contain in himself’ the end of the same action because he
is not in a position to. He doesn’t know what the real end of the action is, and is therefore
not in a position to make it his own—to choose, freely, to contribute to its realization.”
Christman suggests that a person’s development of some desire D is autonomous only if
the person “would not have resisted that development had [he] attended to the process
[by which it was developed]” (Christman, “Autonomy and Personal History,” p. 11). But
Korsgaard’s point is that the victim of deception or manipulation cannot “attend to the
process” by which her desires are influenced without profoundly altering the process itself.
Later, Christman suggests that the agent’s “ability to reflect on her own mental states and
the forces that affect them” could be exercised before the influences occur, the question
thus being: “Would she resist submitting to these techniques and their effects” if she knew
“before the fact” what they amounted to? (p. 21). This sounds more like Scanlon’s test,
which I discuss later in the article. As I explain, insofar as the question is what sort of
treatment someone could reasonably accept, the issue is no longer essentially: what sort
of treatment undermines the agent’s autonomy?
Buss Valuing Autonomy and Respecting Persons 219
45. Again, it is, in principle, possible for us to become aware of any given influence
on our reasoning, even though we cannot become aware of all of them. But this possibility
does not distinguish nonintentional influences from the intentional variety. No matter
what the origin of a nonrational influence, its effectiveness as an influence depends on
the extent to which it is not a constituent of the agent’s practical reasoning. Even an
unintended influence will tend to lose its power when it is called to the agent’s attention.
Conversely, even the intended influences of a manipulator can retain their power when
the agent becomes aware of them. Indeed, their power can sometimes increase as a result.
(This may be what happens in the sort of case Sartre describes in order to illustrate one
sort of bad faith: the woman who leaves her hand in the grasp of the man who has taken
it is in the not terribly unusual situation of allowing her emotions to be “worked on”
nonrationally by another without admitting to herself that this is what she is doing [Jean-
Paul Sartre, Being and Nothingness, trans. Hazel Barnes (New York: Simon & Schuster, 1956),
pp. 97–98]. Mark Miller describes a very different way in which the victims of manipulation
can be witting victims: the success of a particular Pepsi commercial, he notes, depended
on the fact that it appealed to the viewers’ sense of superiority at being wise to the fact
that Pepsi was trying to manipulate them into buying Pepsi [Mark C. Miller, “Deride and
220 Ethics January 2005
no different from the aesthete seducer. More importantly, like the aesth-
ete seducer, most ordinary seducers aim to influence their beloved’s
choices in extrarational ways, without calling attention to what they are
up to. At some point, their intentions will become clear. But not, they
hope, before the beloved is already well primed to encourage them.
It is, of course, impossible to prove that seductions play an essential
role in many good lives. But it does seem highly plausible that a certain
amount of deception—about what one is up to, and about what sort of
person one is, really—and a certain degree of manipulation—to the
end of turning the beloved into one’s lover—is a necessary condition
for the possibility of (at least most cases of) mutual romantic love. The
necessity enters in two different places. First, for many of us, transform-
ing the beloved’s attitude by nonrational means, some of which involve
a certain amount of falsification, is integral to our way of expressing
our love to someone who does not love us back—yet. As Sartre says, we
want to fascinate the beloved, to captivate him, to inspire him—not so
as to convince him to deliberately choose to love us (for all the right
reasons, of course!)—but so as to induce him to “fall in love” with us.
At a minimum, we try to “be on our best behavior,” to “hide our warts.”
Usually, we go further: we become atypically animated, or grave; we act
so much more interested in his tales about his eccentric uncle with the
talking parrot than we ever would have been before—or ever will be
again. We do not deliberately and self-consciously “act a part.” But the
part is a bit of an “act” nonetheless. Not in every case, of course. But,
again, in most of them.
The second reason why manipulation and deception are a necessary
part of most early romances—and so a necessary condition for the pos-
sibility of the mutual love into which these romances evolve when things
go well—is that most early romantic love is too fragile to withstand
complete straightforwardness. A lover is not likely to get very far with
her beloved if she marches up to him and announces: “I really love you.
Would you please (please, please, please) love me back?” In order to
secure her beloved’s “contribution to [her] end,” she cannot simply
“put the fact before him and ask for his contribution.”47 On the other
hand, she has to do something or he won’t even know she’s alive. So,
she has to interact with him in a way that does not reveal the true nature
of her ends. She has to keep her “underlying maxim” “obscure.” (What
she must do, moreover, if mutual love is to have a chance is precisely
what her instinct for self-preservation naturally inclines her to do: since
she is not sure how she would survive his rejection of her intentions—
since his rejection of her intentions would, she knows, feel like a rejec-
that they could not reasonably reject insofar as they share this con-
cern”;50 “the most general moral demand is that we exercise our capacity
for self-governance in ways that others could reasonably be expected to
authorize.”51
It is hard to argue with this. But this is because it offers us so little
to argue about. In particular, it takes no stand on two issues that are
central to any attempt to ground morality in autonomy. First, it is silent
about whether “authorizations” that are (partly) determined by manip-
ulation and/or deceit can themselves count as “self-governed.” Second,
it is silent about what a self-governing agent might authorize. In short,
on this interpretation of the moral value of autonomy, the moral value
of autonomy is not the basis of any substantive moral conclusions. In
particular, it provides no ground for thinking that people would be
unreasonable to endorse manipulation and deceit in a wide range of
cases. If this would, in fact, be unreasonable, the explanation lies
elsewhere.52
According to Scanlon, reasonable human beings would regard it
as unacceptable for one person to “bombard” another with “one-sided
information” or to “create an environment in which [the other person]
is distracted from certain questions by . . . competing stimuli.”53 But if
I am right, then Scanlon is wrong—or, anyway, misleading: reasonable
human beings would permit such behavior, not only in “emergency
cases,” or “threat cases,” or “paternalistic cases,”54 but also in cases where
there are no countervailing moral pressures. They would permit (and
even welcome) such behavior under the circumstances of ordinary
courtship and under many other circumstances as well. They would
judge it permissible—they do judge it permissible—for someone to use
her tone of voice, the volume at which she speaks, her facial expression,
her body language, to captivate or distract another rational agent so as
to elicit from him the sort of response she wants—whether this response
is an emotion, like love, or an act of forgiveness, a determination not
your emotions without your acknowledging that this was the sort of
situation in which such manipulation was permissible, then this is not
because you have not actually given your consent but because in this
context it would be reasonable for you to object to such nonstraight-
forward behavior.56
Again, disagreements about these matters should not distract us
from the main point which they help to illustrate, namely, that the
substantive constraint on what counts as treating persons as ends is
provided, not by the moral value of rational agency, reasonable choice,
or autonomy (the disputants can all concede that it would be wrong to
permit behavior that most rational agents could not reasonably en-
dorse), but by a particular conception of what counts as “reasonably
favorable conditions” for choice. Unlike Kierkegaard’s aesthetes, I do
not think that aesthetic standards have priority in determining what is
“reasonable.” But nor do I believe that most reasonable people would,
as Scanlon suggests, “feel alienated” and upset if they were to discover
that the reason they had fallen in love with someone,57 or done any
number of other things as a result of interacting with her, was because
she had failed to disclose her motives and intentions and thereby failed
to conform to the “ideal of straightforwardness in human relations.”
Of course, we don’t want to be lied to. My point is that if, in the
circumstances, lying is wrong, this is not for the reasons suggested by
talk of “environments” “created” by “one-sided bombardments” and
“distractions.”
Taken together, the reflections of this section call attention to how
many human interactions involve intentional and unintentional influ-
ences that do not depend on the exchange of considerations for or
against some response. We are reminded, too, of how many such in-
teractions are mutually respectful and even seem to have a value beyond
price. To be sure, most of the time most of us prefer to know what other
people intend with respect to us. But there are limits—reasonable
limits—to our strong interest in encountering reality “straightforwardly.”
56. Douglas Husak has made a similar point in challenging the relevance of autonomy
to the moral status of paternalism: “‘When is it reasonable to believe that a person would
consent to an interference?’ The only sensible answer to this question, though one that
may not be particularly illuminating, is that it is reasonable to believe that a person would
consent to whatever interferences are reasonable. The proper focus, then, is on whether
an interference is reasonable—not on whether a person would consent to it. The notion
of consent is retained in the criteria to create the impression that a concern for moral
autonomy is preserved. In reality, however, consent plays only a token role in the criteria.
Once the notion of consent is omitted from the criteria and it becomes evident that the
central focus is on the reasonableness of the interference, it is easier to appreciate that
the attempt to employ the notion of moral autonomy as a barrier to justify paternalism
is unsuccessful” (Husak, “Paternalism and Autonomy,” pp. 34–35).
57. Scanlon, “The Significance of Choice,” p. 157.
226 Ethics January 2005
58. This includes wanting to know the reasons we have for not giving pride of place
to aesthetic reasons. It is important to stress this point, since as Elijah Millgram has noted
in correspondence, someone (e.g., the Seducer) can put someone else (e.g., Cordelia)
“in touch with lots of reality (showing her the seamy underside of everything, and so on),
with the aim of making a cynic of her (for aesthetic reasons).” Of course, the Seducer
does not think that there are any overriding nonaesthetic reasons. But his biggest mistake
from the moral point of view is that he dismisses as irrelevant what other people—most
importantly, Cordelia—take to be normatively significant.
59. Having rejected the hypothesis that subliminal advertising is bad because it is
subliminal (see n. 32), Scanlon then speculates that it is objectionable (at least if it works)
because “it causes us to act . . . by making us think we have a good reason for so acting,
even though we probably have no such reason” (Scanlon, “Freedom of Expression,” p.
525). Surely, there is more to it than this. After all, there are circumstances under which
we can reasonably object to having been manipulated into doing the very thing we have
most reason to do, and this is because under these circumstances we can reasonably object
to having been kept in the dark about some relevant facts. What’s more, the case of the
Seducer suggests that it is usually relevant not only whether the one “causing” us to do
something actually thinks that we have good reason to act this way but also whether it is
reasonable for him to think this.
228 Ethics January 2005
consult with, learn from, or defer to, the verdicts reached, or the ends
set, by other rational agents.65
If manipulation and deceit are sometimes morally acceptable, then
this is because the valuable conditions I have just identified are not the
only valuable conditions, and because the degree to which these con-
ditions are compromised makes all the moral difference in the world.
A conscientious moral agent has her hands full maintaining the right
balance, and this is never truer than when she wants someone to love
her. Because falling in love is intrinsically valuable, the deception and
manipulation integral to (most) early courtships is not only a necessary
condition for the possibility of something else of great value (mutual
love); it is not only a necessary means to a valuable end. It is also a part
of something good in itself—or at least it can be. At the same time,
however, there is a price to pay for this good: the vulnerability that
comes with allowing another person to determine, and even control,
not one’s decisions but one’s emotional life.
To be captivated by the beautiful and the interesting is to become
vulnerable to their charms. It is to be possessed, and so, to be less “self-
possessed,” and so, to have less control—not, again, over one’s decisions
but over one’s happiness. When we read poetry, or go to the theater
or the concert hall, we consent to enter this vulnerable condition; and
the seductions to which we expose ourselves are pretty strictly delimited.
Even so, we can become the victims of demagoguery: the sights and
sounds of Nuremberg can sweep us far away from a good human life.
The risks are even greater when we have not consented to be played
upon, and when the time and place of the seductive influence are not
strictly delimited. This is why it is important that those who move us to
love them not do so in the role of the self-conscious artist who has no
concern for our happiness.
This brings us back to Johannes and Cordelia. Johannes’s moral
fault is not, we have seen, that he fails to recognize and appreciate the
unconditional value of the human capacity for self-government. And it
is not that he manipulates and deceives another human being, where
these forms of human interaction are conceived in purely nonmoral
terms. I have suggested that his behavior is morally criticizable because
he manipulates and deceives another human being with the intention
of submerging her narrative in his—because he has no intention what-
soever of relating to her as an equal, no intention of interacting with
her in a world which belongs to her to the same extent that it belongs
65. As many have noted, the Categorical Imperative appears to tell us not to make
an exception of ourselves. If this is right, and if the Categorical Imperative is the law of
autonomous agents, then to be autonomous is to regard other autonomous agents as
equals.
Buss Valuing Autonomy and Respecting Persons 231
66. Scanlon himself calls attention to the irreducible importance of concern for
another’s welfare. “One can,” he claims, “reasonably refuse to grant others license to
ignore the costs of the expectations they lead one to form.” This is why a reasonable
person would endorse the following Principle L: “If one has intentionally or negligently
led someone to expect that one is going to follow a certain course of action, X, and one
has good reason to believe that that person will suffer significant loss as a result of this
expectation if one does not follow X, then one must take reasonable steps to prevent that
loss” (Scanlon, What We Owe to Each Other, pp. 300–301). More generally, Scanlon notes
that “a person can reasonably reject a principle if 1) general acceptance of that principle
in a world like the one we are familiar with would cause that person serious hardship,
and 2) there are alternative principles, the general acceptance of which would not entail
comparable burdens for anyone.” Thus, “in order . . . to decide whether a given principle
can reasonably be rejected we will need some interpretation of the terms ‘serious hardship’
and ‘comparable burden’” (T. M. Scanlon, “Value, Desire, and the Quality of Life,” in The
Quality of Life, ed. Martha C. Nussbaum and Amartya Sen [Oxford: Oxford University Press,
1993], pp. 185–207, p. 196).
67. It is worth stressing here that, from the moral point of view, the aesthete’s problem
is not simply that he overvalues autonomy—that he gives it too much weight, relative to
other valuable things. Even if he values other things, too (e.g., human emotions, the
capacity to feel pain), he necessarily values them aesthetically and so does not attribute
to them the proper (moral) significance. I have Elijah Millgram to thank for helping me
to see this point.
68. O’Neill, “Between Consenting Adults,” p. 119.
232 Ethics January 2005
69. “If I yield myself completely to another and obtain the person of the other in
return, I win myself back; I have given myself up as the property of another, but in turn
I take that other as my property, and so win myself back again in winning the person
whose property I have become” (Immanuel Kant, Lectures on Ethics, trans. Louis Infield
[New York: Harper Torchbooks, 1963], p. 167, cited in Barbara Herman, “Could It Be
Worth Thinking about Kant on Sex and Marriage?” in A Mind of One’s Own: Feminist Essays
on Reason and Objectivity, ed. Louise M. Anthony and Charlotte Witt [Boulder, Colo.: West-
view, 1993], pp. 53–72, p. 60).
70. Herman, “Could It Be Worth Thinking about Kant on Sex and Marriage?” p. 63.
Of course, in ideal cases, marriage is something to which both parties consent; and in so
doing, they acknowledge certain obligations to one another. In contrast, in the early stages
of a romance, the two parties do not (and are not in a position to) consent to promoting/
protecting one another’s welfare. The lover’s commitment to his beloved’s happiness is
(and initially must be) something he keeps to himself; and the beloved will generally have
no special commitment of any kind to the lover.
Buss Valuing Autonomy and Respecting Persons 233
***
If in most situations we are morally required to inform others of
the ends we are trying to get them to endorse, and of our reasons for
wanting them to endorse these ends, this must be because they could
not reasonably expect us to treat them otherwise in these situations.
But if they could not reasonably expect us to treat them otherwise, this
cannot be because there are independent (metaphysical) reasons for
thinking that such treatment prevents them from exercising their ca-
pacity for autonomous choice. And whatever the nonmetaphysical rea-
sons may be for rejecting such treatment, they are not always decisive:
most reasonable people do not expect to be informed at all times about
how others are trying to get them to think; to the contrary, they expect
not to be informed in many cases; indeed, they do not always want to
be treated “straightforwardly.”
Because it is very difficult to deprive someone of her ability to
govern herself (difficult, i.e., without also depriving her of the ability
to think), and because the requirements for treating someone with
respect cannot be read off of the fact that she can govern herself, the
moral status of manipulation and deceit is far more complex than has
commonly been supposed. There are many manipulative acts which
rational agents cannot reasonably endorse. But this is not because such
manipulation “violates the autonomy” of the one manipulated. Deceit
and manipulation can be motivated by a genuine (nonmoral) appre-
ciation of the unconditional value of the capacity to govern oneself.
And there is no credible conception of autonomy according to which
71. The case of pedagogy is an interesting one. Is what Johannes does to Cordelia
worse than what a teacher does to a student when she manipulates him into gaining
philosophical insight, with the knowledge that this will shatter his secure world and
may well leave him terribly unhappy? Christman suggests that the moral acceptability
of a “pedagogical technique” depends on whether the students are able to “reflect and
critically appraise the educational methods and principles to which they are being
subjected” (Christman, “Autonomy and Personal History,” p. 20). This is surely true with
respect to some such techniques. But in many, if not most, cases students will not be in
a good position to assess their teachers’ methods until these methods have had their effect.
What’s more, the sort of nonrational influences that come to my mind when I think about
teaching are not even really “techniques”; and they are not the sort of “methods” into
whose acceptability one can gain much insight by “critically appraising” them. They consist
in such things as body posture, facial expression, tone of voice, gesture—elements of the
presentation of self of which a good teacher is not even conscious but by means of which
she manages to present the material in a way that makes (yes, “makes” is the right word
here) at least some of the students think that they are being told, or shown, something
that may actually be worth caring about.
234 Ethics January 2005