Sex, Culture, and Justice: The Limits of Choice (33 page)

Feinberg,
Harm to Self,
172.

for that purpose, and so their manufacturers and retailers are not com- plicit in the act. So, if there is a paternalistic prohibition of suicide in two-party cases only, able-bodied people will be able to commit suicide without falling foul of the law. On the other hand, someone who is seriously disabled—perhaps she is paralyzed from the neck down and so can communicate her wishes but not carry them out—is unable to commit suicide without assistance. If a distinction is drawn for the purposes of the law between single-party and two-party acts, the dis- abled person cannot commit suicide whereas the able-bodied person can, and this seems unjust since it undermines basic equality.

Note, however, that a distinction between single- and two-party pa- ternalism is not necessary for this inequality of access to suicide. Such an inequality also occurs if suicide is completely illegal, whether as- sisted or not. Under these conditions an able-bodied person contem- plating suicide is unlikely to fear legal sanction (since they expect to be dead and so not subject to prosecution), but it is likely that the disabled person will find it difficult to obtain a willing assistant (since the assis- tant will remain alive and thus vulnerable to the law). Such an inequal- ity of access is true of all illegal acts. If an able-bodied person wishes to rob a bank, she can (attempt to) do so alone. However, if a seriously disabled or paralyzed person wishes to do so, she will have to find an assistant, and the fact that bank robbing is illegal may make this diffi- cult. In other words, if actions are justly prohibited, it is not an objec- tion that this makes them hard to perform, or that it is harder for some people than others. Bank robbing is harder for those who do not have a gun than for those who do, but this is of no normative consequence. The fact that there is an inequality of access to suicide even where no distinction is made between single- and two-party paternalism sug- gests that the problem lies not in the distinction but in the very prohibi- tion. Readers who feel that suicide should be illegal even for those peo- ple who act alone should not be overly troubled by the issue of unequal access, since as I have shown, it applies to all laws. If, on the other hand, suicide is justly permitted because there is an important human interest in being able to end one’s life, it follows that assisted suicide ought to be permitted as well. I have not claimed that there ought to be a criminal prohibition against every potentially harmful act involving two parties. There are many harms, either potential or actual, that one party can justly inflict on other with her consent, such as risky surgery that may extend the patient’s life but may kill her, or surgery (such as

the amputation of a limb) that will deprive the patient of an important capacity but is necessary for her to survive. I have not argued that such acts should be prohibited, because in these cases the harms involved are outweighed by the benefits, and the benefits are not dependent on the existence of an unjust social norm. Instead, I have argued that there is a case for prohibition when a serious harm is counterbalanced only by benefits that rely on social norms. If there is a good argument in favor of permitting a practice, it should be permitted for all, whether assistance is need or not. However, where a practice merits prohibition, a useful guide is that the prohibition should apply to the person who inflicts the harm rather than the person who suffers it.

Hard versus Soft Paternalism

The last of Feinberg’s distinctions to be considered is between hard and soft paternalism. He writes: ‘‘Soft paternalism holds that the state has the right to prevent self-regarding harmful conduct (so far as it
looks
‘paternalistic’)
when but only when
that conduct is substantially nonvoluntary, or when temporary intervention is necessary to establish whether it is voluntary or not.’’
39
Feinberg’s contention is that soft pa- ternalism is not paternalism at all. So, if my approach is a species of soft paternalism, once again it will not have to face antipaternalist charges. Despite possible appearances, however, my approach is not a species of soft paternalism.

My approach might appear to fall under the category of proscribing ‘‘substantially nonvoluntary’’ conduct as a result of my arguments about social construction. While it is true that social construction ren- ders conduct
in some sense
nonvoluntary, that sense is not of the sort required for soft paternalism. The most relevant sense is that, as I have argued, people might harm themselves in a way that runs counter to their first-order autonomous choices but in line with their second-order choices: they might make a second-order choice to alienate their first- order autonomy. I proposed that it is misleading to describe such a choice as being fully, or first-order, autonomous. However, this argu- ment does not seem to merit the soft paternalism label, since the non- voluntary conduct that Feinberg and others have in mind involves

Ibid.
,
12; emphasis in the original.

things such as ‘‘ignorance, coercion, derangement, drugs’’
40
rather than the conscious choice to suffer a disadvantage for a greater goal.
41

The second sense in which my approach might appear to be a variety of soft paternalism would involve a misunderstanding. It might be sug- gested that, according to my arguments about the social construction of preferences, all conduct is nonvoluntary because all conduct is the result of socially formed preferences. If, as Cass Sunstein argues, ‘‘au- tonomy should refer . . . to decisions reached . . . without illegitimate or excessive constraints on the process of preference formation,’’
42
and if social construction were to constitute such a constraint on preference formation, my approach might seem to legitimate
any
restrictions on
all
preferences, on the grounds that
no
preferences are autonomously reached. Phrasing my argument in this way would be to invoke a night- marish totalitarianism, with the state able to intervene in any and all preferences—far removed from my more limited proposals. Such a formulation would also obscure the fact that, as I argued in Part One, I reject the idea that there is a true subject behind social construction, that behind socially constructed ‘‘nonvoluntary’’ conduct there lies at- omistic, purely ‘‘voluntary’’ conduct. Social construction does not mys- teriously obscure real interests and choices, but is the basic require- ment of choice. As Danny Scoccia points out:

An acceptable theory of autonomous desire should have the consequence that desires based on brainwashing, knee-jerk conformism, and an unreflective response to natural impulses are nonautonomous. . . . On the other hand, such a theory would also have the consequence that few of our desires are or ever could be autonomous, for it is obvious that we (or our wills) cause few if any of our desires. Since ‘‘autonomy’’ in the sense being considered here surely is achievable, an acceptable theory will not say that autonomously formed values are cre- ated by the agent.
43

Thus it is not part of my argument to suggest that paternalism is justi- fied if and because the choice to be overridden is based on faulty rea-

Ibid.

On the other hand, as I discuss in the ‘‘Loyalty’’ section of Chapter 7, this argument may well render my approach nonpaternalist in the same way as Arneson’s dueling example.

Sunstein, ‘‘Preferences and Politics,’’ 11.

Danny Scoccia, ‘‘Paternalism and Respect for Autonomy,’’ 327–28.

soning or irrationality, and it is not part of my approach to suggest that the individual concerned
would
consent to the paternalistic action
if
she were fully rational.
44
As I emphasized in Chapter 5, self-harming practices can be a fully rational response to the prevailing options and social context. If social norms demand that a second-order goal entails a first-order sacrifice, then no failure of rationality is involved. What may be involved is injustice.

This is not to argue that cases of soft paternalism cannot be identi- fied. Clearly, there can be cases where individuals’ decision-making abilities are particularly compromised by the sorts of phenomena that Feinberg describes; and in such cases, I tend to agree with the soft paternalist line. However, I do not seek to justify or excuse my ap- proach by labeling it as a species of soft paternalism, and therefore not really a paternalist at all. Instead, I have sought to argue that proscrip- tion in the circumstances I have outlined is itself a contribution to, perhaps even a requirement of, liberal justice.

To conclude this brief discussion of paternalism: I have argued that my proscriptive proposals may escape the charge of paternalism via Arneson’s definition and his dueling example, but not through the route of soft paternalism. If my proposals are paternalistic, then, the potential evil of their paternalism is mitigated by the fact that my pro- posals are harm-preventing rather than benefit-promoting, so that many options are left intact; and the fact that they should, wherever possible, be applied to the ‘‘other’’ individual, the individual who harms, in two-party cases. These considerations were adduced in this chapter; the most important arguments in favor of my proposals, how- ever, are found in the book as a whole.

Autonomy

Here I consider the extent to which my theory promotes autonomy, and what kind of autonomy it promotes. Some theorists have argued

This proposal is advanced by Dworkin, who argues that ‘‘we would be most likely to consent to paternalism in those instances in which it preserves and enhances for the individ- ual his ability to rationally consider and carry out his own decisions’’ (‘‘Paternalism,’’ 81). Arneson criticizes Dworkin on the grounds that his proposal fails ‘‘to safeguard adequately the right of persons to choose and pursue life plans that deviate from maximal rationality’’ (‘‘Mill versus Paternalism,’’ 474). Because my proposals are not based on rationality, Arne- son’s criticism does not apply to them.

that paternalism is justified if and because it
enhances
autonomy. Thus Dworkin argues that paternalism should be justified by ‘‘a concern not just for the happiness or welfare, in some broad sense, of the individual but rather a concern for the autonomy and freedom of the person.’’
45
Autonomy is enhanced, for Dworkin, because paternalism should be used to satisfy the person’s own rational desires, a formulation that I have rejected.

Similarly, Scoccia sets out three conditions under which paternalis- tic action does not violate autonomy:

Interfering with a person’s choices for his own good does not violate his autonomy if: (
a
) the person has highly autonomous desires, but the choice made does not accurately express those desires, and the person would consent to the interference if he were fully rational, or (
b
) the person has low autonomy desires, and the interference is necessary to preserve his potential to develop it later . . . [or] (
c
) the person has low autonomy desires, the interference would increase the autonomy of his desires (e.g., by removing an obstacle like neurosis or false conscious- ness to autonomous desire formation), and the person would not object to the interference if he were rational and had high autonomy desires.
46

These strategies may be successful. Cases (b) and (c) resemble the soft paternalism just described, which is at least initially plausible (I have not considered it in detail). The most relevant to my approach, though, is case (a). One might argue that paternalism is justified if and when one of the individual’s second-order autonomous goals is thereby sup- ported, so that the individual’s ‘‘highly autonomous desires’’ are better met.

Thus, in the case of Jenna Franklin’s breast implants discussed in Chapter 5, one might argue as follows:

1. Franklin has a second-order autonomous desire to be famous. (This premise is derived from her own statements.)
   

Dworkin, ‘‘Paternalism,’’ 81.

Scoccia, ‘‘Paternalism and Respect for Autonomy,’’ 330–31.

1. If Franklin is to be famous, she must have breast implants. (This premise is also derived from her own statements. For the pur- poses of argument, assume it to be true.)
   

2. Therefore, banning breast implants thwarts Franklin’s second order desire to be famous.
   

3. Franklin has a (separate) second-order autonomous desire to be healthy. (This premise is imputed to her—possibly as a rational being.)
   

4. Breast implants will damage her health. (Arguments supporting this premise were given in Chapter 5. For the purposes of argu- ment, assume it to be true.)
   

5. Therefore, allowing breast implants thwarts Franklin’s second- order desire to be healthy.
   

6. Therefore, either banning or allowing breast implants will thwart one of Franklin’s second-order desires.
   

7. Therefore, it is not an argument against banning breast implants that to do so would be to overrule her second-order autonomy.
   

   This argument is valid. However, there are two problems. First, the argument as it stands provides equal support for allowing or proscrib- ing breast implants. It simply states that either will both harm and promote her second-order autonomy. Second, the balance may be tipped against proscription by the following consideration: the antipa- ternalist defender of autonomy might argue that it should be
   Franklin,
   not the state, who decides which of her second-order goals—health or fame—to choose. (Such an argument might invoke some idea of third- order autonomy, according to which one chooses which second-order goal to follow.)
   

   We can respond to these two problems by highlighting another plau- sible feature of the case, which I posited earlier:
   

8. Franklin would prefer to be famous without breast implants, if that were possible.
   

   If this premise is added, the balance is tipped in favor of prohibition. For, as I have argued, banning breast implants will have the beneficial effect that it cannot, by definition, be a requirement of fame that one have breast implants (for no one would have them and so, if the re-
   

   quirement persisted, no one could be famous).
   47
   It follows, then, that only with prohibition will Franklin be able to achieve both her second- order goals of fame and health. As discussed earlier, premise 9 means that my proposals would not be paternalistic on Arneson’s definition; the purpose here is to show that they might also be construed as auton- omy-maximizing.
   

   It will not be a good general strategy, however, to attempt to justify my proposals on the grounds that they are autonomy-maximizing. There are three potential problems. First, there might be people for whom (the equivalent of) premise 4 does not apply. Such people can be described as having unusual—we might want to say irrational— second-order goals. Second, there might be people for whom (the equivalent of) premise 9 does not apply. These people can be described as having unusual or irrational first-order goals. In either case, pater- nalist proscription may well reduce their second- or first-order auton- omy. Third, however, autonomy-maximization is a bad goal. In the rest of this section, I consider each of these considerations in turn.
   

   ‘‘Irrational’’ or Unusual Second-Order Goals
   

   The first case to be considered is those people who have unusual or irrational second-order goals. In the breast implant example, we might describe these people as those who have no desire to be in good health. Since this condition seems implausible and, almost uncontroversially, irrational, it is better to describe them more sympathetically as people who are able to rank their second-order goals and for whom health has a lower priority than another of their second-order goals (in the Frank- lin case, fame). To describe them even more sympathetically (and thus in a way that poses even more problems for my argument), we could
   

It is possible, of course, that it might still be a requirement of fame for a woman that she has large breasts, so that prohibition of breast implants prevents small-breasted women from achieving fame. This seems unlikely, for several reasons already detailed (for example, that breast implants create as much as respond to a widespread desire for large-breasted women). However, if the requirement of large breasts were to remain for some particular careers (such as some sorts of topless or underwear modeling), it seems no more problematic that small-breasted women would not have access to such careers than it seems problematic that clumsy people cannot be brain surgeons—or at least, if the large-breast requirement is unjust (perhaps because there is no real
need
to have large breasts to perform the relevant functions), the injustice ought to be resolved by removing the requirement rather than by implanting silicone into women’s bodies.

describe them as those for whom the
risk
of bad health is of less impor- tance than the achievement of other goals.

This final, most sympathetic phrasing suggests a possible disanal- ogy between breast implants and
fgm
. The harms of breast implants are, by and large,
risks
rather than certainties, whereas the harms of
fgm
(one might argue) are certain to occur as an intrinsic part of the procedure. Against the argument we might say that for consenting adult women undergoing
fgm
in clinical conditions many of the harms of
fgm
(infection, death) become risks, and other harms (the loss of the clitoris) might be described as the very purpose of the proce- dure and not as harmful for the women involved. Moreover, the Insti- tute of Medicine found that many of the side effects of breast implants, such as capsular contracture, are more or less certain to occur over the course of an implanted woman’s life. However, in favor of the argu- ment it must be stated that the removal of the clitoris remains harmful in the sense of destroying a bodily function (sexual pleasure), and that infibulation brings with it other certain harms which are not the pur- pose of the procedure (hindering urination, menstruation, and child- birth; increasing the risk of infection; and so on).

My proposals must, then, respond to the challenge that for some people, the disvalue of the risk of bad health is outweighed by other second-order autonomous goals. Indeed, this seems to be true of all of us who willingly drive a car, eat fast food, and so on. Dworkin takes an uncompromising line with many such people: ‘‘Consider a person who knows the statistical data on the probability of being injured when not wearing seat belts in an automobile and knows the types and gravity of the various injuries. He also insists that the inconvenience attached to fastening the belt every time he gets in and out of the car outweighs for him the possible risks to himself. I am inclined in this case to think that such a weighing is irrational.’’
48
In this example, however, the charge of irrationality relies on the imbalance between the tiny cost of fastening one’s seat belt and the huge potential damage that can be done to a person in a car accident. It may not be possible to make such an overwhelmingly clear judgment in other cases. Nonetheless, this balancing act—what Feinberg calls the ‘‘balancing strategy’’
49
—is pre-

Dworkin, ‘‘Paternalism,’’ 76–77.

‘‘To say that the need to protect people from their own foolishness is always
a
‘good and relevant reason’ for coercive legislation, is not to say that it is in any given case a decisive reason. Rather, it leaves open the possibility that in that case reasons of a quite different kind weigh on the other side, and that those other reasons (including respect for personal auton-

cisely what is called for in my proposals. It will be necessary, with each individual practice, to consider its overall harms and risks against the weight of the competing considerations, so as to discern where the balance of reasons lie. In this respect my proposals are no different from any other theory of justified paternalism.

Where they are different, however, is in their claim that one of the considerations that must be taken into account is the fact (if it is a fact) that the benefits of the practice result only from a social norm. If the practice is worth following only because a social norm decrees that it regulates access to some benefit, that fact should, I have argued, weigh very strongly against the practice in question. I have also argued that both physical
and
status harm ought to be placed on the scales, in favor of proscription.

That having been said, I do not wish to argue that this approach can never undermine individuals’ autonomy. It would undermine a person’s autonomy in the case where proscription serves no second- order goal of hers, or serves only a goal which she ranks as less impor- tant, but does prevent her from achieving her actual, or more impor- tant, second-order goal. Thus if a woman (1) wanted breast implants in order to be famous,
and
(2) ranked the risks to her health below her desire to be famous,
and
(3) would become famous with breast im- plants,
and
(4) would not become famous
if and because
she was unable to have breast implants, then proscribing breast implants
would
under- mine her second-order autonomy. This undermining would, in turn, be one objection to proscription. Taken in the context of my arguments throughout this book, and the extreme particularity of the case, the objection is not decisive. However, it ought to be acknowledged; I take seriously Isaiah Berlin’s point:

It is one thing to say that I may be coerced for my own good which I am too blind to see: this may, on occasion, be for my benefit; indeed it may enlarge the scope of my liberty. It is another to say that if it is my good, then I am not being co- erced, for I have willed it, whether I know this or not, and am free (or ‘‘truly’’ free) even while my poor earthly body and fool-

omy) may in the circumstances have still greater weight. Thus, it is possible to defend legal paternalism, as we have defined it, while arguing against paternalistic legislation in particular cases. We can call this approach ‘the balancing strategy.’ The anti-paternalist has a heavier argumentative load to carry’’ (Feinberg,
Harm to Self
, 25; emphasis in the original).

ish mind bitterly reject it, and struggle against those who seek however benevolently to impose it, with the greatest despera- tion.
50

‘‘Irrational’’ or Unusual First-Order Goals

The second case to be considered is those people who have unusual or irrational first-order goals. In the breast implant example, we might describe these people as women who want breast implants in and of themselves. Regardless of social norms or any other second-order goal, they just want breast implants.

I dealt with this issue much earlier in the book. Following Foucault, I argued that it would be impossible for someone to want to undergo a practice
regardless
of social norms—unless, perhaps, that desire were one that made no sense in the context of social norms, such as cos- metic knee implants. What would be possible, however, would be for a woman actively to consider a practice that makes sense only in the context of social norms (such as breast implants) and make an autono- mous decision to conform to the social norm. However, for such a woman to fit into the category of someone with an unusual first-order desire rather than a second-order desire, that woman must
either
want to have breast implants in themselves, not as a means to a higher second-order goal (such as beauty or self-esteem),
or
prefer breast im- plants above all other alternatives (counseling, feminist consciousness- raising, a padded bra) as a means to that higher goal. The first possibil- ity is at least odd and at most implausible since, by definition, such a woman could provide us with no reason for her desire. The second is more plausible. Many of the women in Kathy Davis’s study of cosmetic surgery felt that cosmetic surgery was their only option for acquiring self-esteem, since alternatives had failed.
51
For these women, however, cosmetic surgery is not their
preferred
option so much as (in their eyes) their
only
option. If other options were available (and proscription might increase the effectiveness of other options by lessening the force

Berlin, ‘‘Two Concepts of Liberty,’’ 25. I do not attempt to define either autonomy or paternalism in such a way as to make it true by definition that my proposals either promote autonomy or are not paternalistic, as Scoccia and Arneson respectively seem to do. Arneson’s strategy for dealing with paternalism seems to be to decide what sorts of paternalism are justified, and then to construct a definition of paternalism that excludes them. The result is that the phrase ‘‘paternalistically justifiable’’ becomes oxymoronic.

Davis,
Reshaping the Female Body.

of certain appearance norms), it seems likely that these women might prefer it.

Nonetheless, once again, it is the case that
if
a woman does in fact want breast implants in and of themselves, or as her preferred means to a second-order goal, then proscribing breast implants
will
hinder her first-order autonomy. Again, this consideration is an objection to my proposals that, though indecisive, should be acknowledged.

The Limitations of Autonomy

I have accepted, then, that under certain circumstances for certain indi- viduals my proposals will limit either first- or second-order autonomy. I have also accepted that this limitation is a relevant objection to my proposals. It is not, however, a decisive objection, because there is no reason to think that autonomy is the only value, or that it is the most important value in a way that rules out trading it off for other values, or that it ought to be maximized.

My arguments hitherto are supported by Steven Wall’s arguments against the idea of autonomy-maximization. Wall argues that auton- omy is a ‘‘distinctive character ideal’’ that has ‘‘intrinsic and instrumen- tal value’’ and should have a ‘‘privileged position in a sound account of political morality.’’
52
These claims notwithstanding, Wall argues that autonomy does not justify antiperfectionism (which would scupper my approach) because autonomy-maximization is not a coherent goal. He points out that autonomy simply cannot be maximized, because it is measured along several dimensions, and maximizing autonomy on one dimension may conflict with maximizing it on another. The di- mensions that Wall cites are, first, the number of autonomous people in a society and, second, the amount of autonomy each one has. To these dimensions we might add that a person’s first-order autonomy can conflict with her second-order autonomy (as in the case of a woman who wants to remain in her cultural community but not un- dergo
fgm
), and that two second-order autonomous goals may conflict (fame and health in the breast implants example). If autonomy con- tains conflicting dimensions, it cannot be maximized.

Even if it were possible to maximize autonomy, though, the most important objection to the idea of autonomy-maximization is that au-