The value of autonomy in medical ethics

Beauchamp and Childress (2001, pp. 185–187) also maintain that “under some conditions a narrow range of strongly paternalistic acts is justified.” This means that in the majority of cases autonomy can override wellbeing. And Beauchamp and Childress (2001, p. 186) are inclined to accept that even when paternalism is justified, the paternalistic action should not “substantially restrict autonomy.”

See, e.g., Beauchamp and Childress (2001, pp. 185– 187) and Glover (1977, p. 75 cf. p. 81). All philosophers accepting the view that patients should be allowed to make choices that these patients admit to be bad, or at least not as good as they could be, do not specify whether they think there to be some cases where autonomous patients’ self-regarding actions can be restricted.

Harris would seem to hold this kind of view. Harris (2003, p. 11) writes as follows: “Concern for welfare ceases to be legitimate at the point at which, so far from being productive of autonomy, so far from enabling the individual to create her own life, it operates to frustrate the individual’s own attempts to create her own life herself.”

Alternatively, it can be taken that autonomy is a property of decisions and choices (see, e.g., Beauchamp and Childress (2001, p. 58)). To the extent that it is necessary from the point of view of the main argument of this paper, I will return to this distinction between autonomy as a property of decisions and choices and autonomy as a property of persons below.

For a defense of a partly substantive conception of autonomy see Oshana (1998).

The views that are referred to as Kantian can be even radically dissimilar to each other, see, e.g., Secker (1999) for discussion.

The most common criticism of the procedural theories of individual autonomy is that in stressing the value of a person’s own independent decisions, they are incompatible with such communitarian virtues as empathy, loyalty, and care and respect for tradition, ritual, and the counsel of others. For a plausible reply to this criticism see Oshana (2001) (whose own theory of autonomy contains substantive elements) and also, e.g., Gylling (2004). Oshana argues that this communitarian criticism of individual autonomy is an objection to a straw man, while Gylling warns us about uncritical acceptance of traditional values. Gylling writes, “If moral integrity is to have any meaning, ethically attentive individuals should critically assess the merits and dismerits of predominant credos. Conservative and unquestioning acceptance of traditional values may turn moral life into a mechanical — even if bona fide - acceptance of either ruthless and inhuman morals or a hollow etiquette, a code of honor.” Gylling (2004, p. 46).

In addition to the philosophers who maintain that patients’ autonomy has value beyond its instrumental value in promoting the patients’ wellbeing mentioned above see, e.g., Mappes and Zembaty (1991), Matthews (2000), Secker (1999), and Young (1998). For discussion of Kant’s and Kantian concepts of autonomy in medicine and some persuasive arguments against the view that medical ethics should refer to these kinds of conceptions to autonomy, see, e.g., Matthews (2000), Secker (1999), and Takala and Häyry (2000).

That I will not argue for the possibility of intrinsic value in general should not beg the question here, since I will be arguing against the view that patients’ autonomy has intrinsic value.

It could be objected that if the proponents of the view that patients’ autonomy has value for the patients over and beyond its value in promoting their wellbeing do not specify how they perceive the value of autonomy, it is implausible to maintain that they consider autonomy to have intrinsic value for the patients. However, at least when other things are being equal, the most plausible interpretation of the view that patients’ autonomy has value for them over and beyond its instrumental value in enhancing their wellbeing is that autonomy has intrinsic, and not economic, aesthetic, etc., value for them.

Sometimes defenses of personal autonomy refer to Kant and Mill (see, e.g., Beauchamp and Childress (2001; pp. 63-64, cf. p. 351)). But since Kant deems the choices a person makes on the basis of her own desires, inclinations, and wishes as heteronomous, Kant is not directly relevant to arguing for the kind of autonomy that has a prominent role in contemporary medicine. For detailed arguments to this effect see, e.g., Matthews (2000) and Secker (1999). And it is unclear to what extent Mill considered autonomy to have intrinsic value, and whether or not putting intrinsic value on autonomy is consistent with his utilitarianism. Be that as it may, Mill did maintain that making our own choices exercises our rational capacities and helps us to build a character (Mill, 1985, pp. 122–124), and these reasons for putting intrinsic value on autonomy are discussed below.

Gillon (2003, p. 310) says that autonomy is morally precious because it makes morality possible. It is not altogether clear how this should be taken. If it means that autonomy should be valued because it gives persons the possibility to perform morally good actions, it can be responded that these goods are overweighed by the grave moral wrongs that autonomy makes possible to commit. Procedural autonomy in itself is neutral between morally good and bad actions and can be used in good and bad ways (see, e.g., Dworkin (1988)). And even if we concentrated on the morally good things that autonomy makes possible, it can be asked are these things better than the kinds of goods that do not involve autonomy? In other words, this line of thinking would just take us back to the question of does autonomy have intrinsic value. Furthermore, we can reasonably consider persons as morally culpable for actions that they did not perform autonomously (see, e.g., Oshana (2002)). For example, consider the case in which a heavily drunk person hits another person with her car with the consequence that the latter person dies. It is intuitively plausible to consider the drunk as morally culpable for the other person’s death even if she did not autonomously choose to kill him. And finally, the view that autonomy is morally precious because it makes morality possible would seem to make the value of autonomy derivative from the value of morality.

Glover (1977, p. 75) also says that if someone wants to start taking heroin, he thinks it right to stop the person, because he gives less weight to autonomy in this matter than to sparing the person “the appalling suffering involved in the slow death of a heroin addict.” Glover thus accepts that autonomy can be restricted only if there is threat of extremely severe harm to wellbeing.

Our lives can of course be unique in the trivial sense that no one else was born from the same parents at exactly the same time, but this would not seem to support the argument for autonomy.

There admittedly are persons who declare that even though the choices they have made were bad, were they given the possibility to go back in time and start all over again, they would make exactly the same choices as they did the first time around. But, as such, their cases do not imply that autonomous persons would really want to repeat choices that they consider to be altogether bad. There are advantages in being able to retain a positive image of oneself in the eyes of oneself and others, and to the extent that these declarations are not these persons’ attempts to present themselves as more competent agents than they have been, it is plausible, I think, that they consider that nobody could have been able to help them and that as a result of making their mistakes they have learnt valuable lessons and consequently know how to avoid similar mistakes in the future.

I do not mean to confuse considering the thoughtexperiment with actually choosing whether or not let others choose for one. However, if the former is to be plausible and effective, it must be quite reminiscent of the latter, and hence follows this problem.

Relevant to this point is, e.g., May’s (2005) response to the criticisms of autonomy maintaining that the conception of self that autonomy usually presupposes cuts the self off from others and therefore cannot account for the significance of the social dimensions of identity, relationships, etc. Instead of considering an autonomous agent to be an island, May maintains that an autonomous person should be seen as a helmsman who steers her own life and may resort to others’ assistance in doing so.

It could be maintained that the problems pertaining to the reasons that allegedly make us reluctant towards letting others choose for us are problems only to those proponents of the argument for autonomy who refer to them and that therefore those philosophers accepting the argument for autonomy who do not speak about these reasons can still be considered to have a plausible view about the value of autonomy. However, of the objections listed above, one, four, and five do apply to these philosophers’ view and these criticisms are sufficient to show that the argument for autonomy that does not refer to the problematic reasons referred to above is not able to establish the view that autonomy has intrinsic value for us.

I now refer to making one’s own choices only, but what I will say below applies, mutatis mutandis, to the case of the Brave New World argument as well.

It could be maintained that although autonomy does not have intrinsic value, it still is a necessary condition of everything that has value for us (see, e.g., Rachels and Ruddick (1989)) and, consequently, it should always be taken into account even if this were not the instrumentally best way to promote a patient’s wellbeing. However, even if we accepted that autonomy conditions the value that things have for us, it seems that then bad choices would be the worse for us the more they were of our own making. In other words, even if we accepted that autonomy has this kind of conditional value, that a bad choice is made autonomously would not make that choice good for one.

I thank Juha Räikkä and two anonymous reviewers for helpful comments on an earlier version of this paper.