Business Ethics And Social Responsibilities 2000 words Please read through all my requirement before accepting. Thanks.

I then evaluate the practical relevance of both concepts of autonomy for use in bioethics.

I argue that it is not appropriate to appeal to either concept toward assessing which patients we ought to respect as autonomous. Finally, I sketch criteria for what I take to be a more promising concept of autonomy for patients.

Key words: autonomy, decision-making, Kant, Kantianism, moral autonomy. I. INTRODUCTION In opposition to the traditional paternalism of the physician-patient rela- tionship, the last few decades have seen patient autonomy emerge as a pivotal concept in human research and health contexts. The doctrine of informed decision-making, the concept of mental competence, and the principles of privacy, veracity, and confidentiality, all reflect the relative- ly recent move to assess, protect and enhance the ability of patients to make autonomous decisions about their health care, by limiting the power of physicians to act in paternalistic ways and by making them more ac- countable to patients. Respect for autonomy has become a fundamental Ñ arguably the paramount Ñ moral principle in bioethics, together with principles of beneficence and justice. Since patients have both a prima facie right to autonomy and a need for care, and since physicians have a duty both to respect the autonomy of, and to provide care for, their pa- Correspondence: Barbara Secker, Ph.D., University of Toronto, Department of Philoso- phy, and Joint Centre for Bioethics; and Rehabilitation Institute of Toronto, Toronto, Ontario, Canada. BARBARA SECKER 44 tients, much bioethical discussion revolves around questions of decisional authority, cast as conflict between the principles of autonomy and benefi- cence.

Proponents of the principle of respect for autonomy need an adequate model of autonomy Ñ a foundation for the right of autonomous patients to make their own health care decisions. Respecting patientsÕ autonomy pre- supposes the presence of autonomy in sufficient degree; however, not all patients are sufficiently autonomous. Thus, we need appropriate criteria in order to help us distinguish those patients who can make decisions auton- omously from those who cannot and whose welfare depends on others deciding in those patientsÕ best interests.

Despite the moral premium placed on autonomy, it is common to find the term used rather intuitively. The concept of autonomy is rarely given much analysis and thus frequently operates in bioethics as a vague, ambig- uous or elusive concept. While several different accounts of autonomy can be identified in the literature, 1 one of the more influential concepts is labelled ÒKantian.Ó 2 This Kantian concept of autonomy and the principle of respect for autonomy find their inspiration in KantÕs deontology, which is one of the major theoretical frameworks underlying bioethical argumen- tation.

Many bioethicists appear to believe that the Kantian notion of autono- my and KantÕs concept of autonomy are one and the same. 3 However, the contemporary Kantian notion has very little at all to do with KantÕs origi- nal. In order to consider the adequacy of each of the contemporary Kantian concept and KantÕs original model for bioethics Ñ and to further bioethics discourse on autonomy Ñ the task of critically distinguishing these often conflated concepts must be undertaken.

Toward these ends, I describe KantÕs model of moral autonomy and the contemporary Kantian conception of autonomy as it is portrayed in bioeth- ics. I then show that these concepts have little in common. Next, I evaluate the practical relevance of the Kantian notion and of KantÕs original con- cept for use in bioethics. I argue that due to moral and practical problems, it is not appropriate to appeal to either concept toward assessing which patients we ought to respect as autonomous. Finally, I sketch criteria for what I take to be a more promising concept of patient autonomy. (Through- out this paper, I refer to KantÕs original concept of moral autonomy as ÒKantÕsÓ concept of autonomy and to contemporary bioethicsÕ concept of autonomy, inspired by Kant, as the ÒKantianÓ concept of autonomy.) KANTÕS CONCEPT OF AUTONOMY 45 II. KANTÕS CONCEPT OF MORAL AUTONOMY 4 KantÕs moral philosophy is built on his resolution of the fundamental problem of human freedom and causal determinism. 5 Laws of freedom and laws of nature are compatible, according to Kant, as long as free will Ñ the ability to act autonomously Ñ is understood to belong to the nou- menal, rather than to the phenomenal, realm. Causality and knowledge are limited to the natural world; and freedom and morality apply to the intelli- gible world. Granted, we cannot know with certainty that freedom exists, but we must necessarily take the idea of free will as a practical and norma- tive postulate if we are to preserve the conception of ourselves as moral agents. Thus, morality presupposes autonomy of the will which, in turn, presupposes freedom. In KantÕs words:

The concept of autonomy is inseparably connected with the idea of freedom and with the former there is inseparably bound the universal principle of morality, which ideally is the ground of all actions of ra- tional beings, just as natural law is the ground of all appearances (FMM, p. 71). 6 Clearly, then, autonomy of the will is the central concept in KantÕs moral philosophy.

Kant describes autonomy of the will as Òthat property of [the rational will] by which it is a law to itself independently of any property of objects of volitionÓ (FMM, p. 59). A being with an autonomous will, then, is one who is self-legislating, rather than subject to pre- or other-given laws. The self-given laws of an autonomous agent are not just any laws, however; these laws are given by the agentÕs pure practical reason, and the form those laws take, according to Kant, is that of the one and only Categorical Imperative Ñ ÒAct only according to that maxim by which you can at the same time will that it should become a universal lawÓ (FMM, p. 39). In KantÕs words: Ò[man] is subject to his own, yet universal legislation, andÉhe is only bound to act in accordance with his own will, which is, however, designed by nature to be a will giving universal lawsÓ (FMM, p.

49).

The Categorical Imperative is the supreme principle of morality Ñ the supreme principle of practical reason Ñ and autonomy, on KantÕs view, is inextricably linked to conformity with this principle. By contrast, heter- onomy results when the Òwill does not give itself the [universal] law,Ó but, rather, Ògoes outside itselfÓ and chooses to act in accordance with the laws of nature or of God Ñ in other words, when it chooses hypothetical imper- BARBARA SECKER 46 atives (FMM, pp. 59Ð60). Furthermore, in contrast to heteronomy or au- tonomy of the wills of finite rational beings, the maxims of a holy will (or an absolutely good will) are necessarily in ÒharmonyÓ with the laws of autonomy (FMM, p. 58).

In order to understand better the connections among autonomy of the will, rationality, freedom, and morality, it is necessary to outline briefly some of the moves made by Kant in section three of his Foundations of the Metaphysics of Morals. First, toward explaining the autonomy of the will, Kant undertakes an analysis of the concept of freedom. He begins with a definition of negative freedom: ÒAs will is a kind of causality of living beings so far as they are rational, freedom would be that property of this causality by which it can be effective independently of foreign causes determining itÓ (FMM, p. 64, my emphasis). Human wills, then, are neces- sarily free in this negative sense Ñ that is, there is no natural necessity of outcome when a rational agent deliberates and chooses to act.

Second, Kant fills out the concept of positive freedom, which goes beyond independence of ÒforeignÓ causes. Since negative freedom alone cannot account for the causality of the will (it only describes what does not determine the will), its causality must be non-alien. Thus, if free will occurs it must be not only negatively, but also positively, free. Its causality must be a capacity for autonomy: ÒWhat else, then, can freedom of the will be but autonomy, i.e., the property of the will to be a law to itself?Ó (FMM, p. 65). So, autonomy presupposes positive freedom.

Third, assuming familiarity with his arguments in section two Ñ where he argues that the supreme principle of practical reason ties autonomy to the capacity to choose universalizable maxims Ñ Kant quickly derives the Categorical Imperative from autonomy. He encapsulates his claims thus far: Òif freedom of the will is presupposed, morality together with its principle follows from it by mere analysis of its conceptÓ (FMM, p. 65). 7 KantÕs next move is to presuppose, since he cannot prove, that freedom must belong to all rational agents, and that Òevery being which cannot act otherwise than under the idea of freedom is thereby really free in a practi- cal respectÓ (FMM, p. 66). Rational beings must regard themselves as free in so far as they, via their practical reason, can attempt to move themselves to action. To this point, then, Kant can make a conditional conclusion along these lines: If we have a rational will, or practical reason, then we are free in a positive sense, and capable of autonomy, and, thus, bound by morality.

Finally, toward the middle of the third section, Kant invokes his distinc- tion between the noumenal and phenomenal realms in order to resolve the problem of freedom and determinism, and to explain how the law of hu- KANTÕS CONCEPT OF AUTONOMY 47 man reason contains the Categorical Imperative. KantÕs view, then, is that all finite rational beings have the capacity for moral autonomyÑfor self, yet universal, legislation. Yet, only those agents who choose to act on the moral law from duty act autonomously and, thus, virtuously. That is, moral autonomy requires a good will Ñ that one acts for the sake of duty and not for any other reason (e.g., self-interest or natural inclination). Those who (choose to) act on some other maxim act heteronomously. 8 III. THE KANTIAN CONCEPT OF AUTONOMY One of the less sketchy descriptions of the Kantian concept of autonomy, as it operates in bioethics, is given by Mappes and DeGrazia (1996) in their introduction to the popular textbook Biomedical Ethics. 9 Following a brief paragraph in which the authors locate the basis of the moral value accorded to individual autonomy in KantÕs view of respect for persons, they ask Ò[b]ut how does Kant understand autonomy?Ó Their response, in its entirety, is:

KantÕs primary focus is on the autonomy of the will. For Kant, ÒAuton- omy of the will is the property the will has of being a law to itself.Ó What Kant calls the Òdignity of man as a rational creatureÓ is due to human beings possessing just that property that enables them to govern their own actions in accordance with rules of their own choosing. Putting aside many complexities in KantÕs own thinking, a Kantian position central in biomedical ethics describes autonomy in terms of self-con- trol, self-direction, or self-governance. The individual capable of act- ing on the basis of effective deliberation, guided by reason, and neither driven by emotions or compulsions nor manipulated or coerced by oth- ers is, on the Kantian position, the model of autonomy (Mappes and DeGrazia, 1996, p. 28, my emphasis).

On this account (and other similar accounts 10) of the Kantian position, then, an autonomous person is a free, independent, self-governing individ- ual. Moreover, an autonomous person decides and acts based on exclu- sively rational reasons Ñ that is, her choices and actions are influenced by reason, not by emotion or inclination. In addition, she is free from physical or psychological constraints, imposed by herself or by others, to choose and act. In sum, a subject who meets the requirements of this Kantian concept of autonomy is self-determining in the sense that she enjoys per- sonal liberty in deciding and acting based on her reason alone. BARBARA SECKER 48 IV. DISTINGUISHING KANTÕS CONCEPT FROM THE KANTIAN CONCEPT The contemporary Kantian view of autonomy in bioethics has little in com- mon with KantÕs original model. In the most general sense, both accounts share the core idea of autonomy as self-governance (of some kind). But KantÕs understanding of this core notion is not self-determination, self-con- trol, self-direction (or any of the other synonyms Kantians use for autono- my). KantÕs ÒautonomyÓ means, literally, rational Òself-legislation,Ó and is necessarily connected with morality. Mappes and DeGrazia describe Kan- tian subjects as governing Òtheir own actions in accordance with rules of their own choosingÓ (Mappes and DeGrazia, 1996, p. 28), 11 whereas, Kant specifies that the self-chosen laws of rational, autonomous persons are uni- versal moral laws. KantÕs conception is not of individual or personal auton- omy, where the central question is ÒWhat do I really want, and is it best for me?Ó; rather, it is of moral autonomy which applies universally, and asks the question ÒIs this what I ought to do?Ó, morally speaking.

Both concepts of autonomy place a premium on rationality. But, as we have seen, for Kant rationality implies something much different than the rational pursuit of desires or preferences with foresight and critical reflec- tion. For Kant, rationality requires logical and volitional consistency such that one could will that the maxims of oneÕs actions become universal laws. Thus, KantÕs understanding of rationality differs significantly from the Kantian view of rationality as merely instrumental and internally co- herent (OÕNeill, 1989, p. 66).

Given that the independence suggested by the Kantian concept may be interpreted as requiring self-sufficiency, 12 we can further distinguish these two concepts of autonomy by noting that Kant is not committed to such a position. As Onora OÕNeill explains:

Kant speaks of beings such as ourselves as finite rational beings not only because their rationality is limited, but because they are finite in many ways. They have limited capacities to act that can be destroyed or undercut in many ways. Self-sufficiency is an incoherent goal for finite rational beings; at most they can coherently aim to minimize their de- pendence on others (OÕNeill, 1989, p. 133, my emphasis). 13 Nothing follows from KantÕs conception of autonomy to the effect that autonomous agents must be independent of or separate from others.

What OÕNeill says of Kantian ethicists in general seems to apply to Kantian bioethicists. She writes: ÒMany of them offer no account of the KANTÕS CONCEPT OF AUTONOMY 49 virtues, or even deny that an account is possible; many treat rights rather than obligations as fundamental; nearly all rely on a preference-based theory of action and an instrumental account of rationality, all of which are incompatible with KantÕs ethicsÓ (OÕNeill, 1991, p. 184). Clearly, Kant and contemporary Kantian bioethicists have very different ideas about agency, rationality, and autonomy. 14 V. ON THE PRACTICAL RELEVANCE OF THE KANTIAN CONCEPT FOR BIOETHICS The moral requirement to respect Kantian autonomy in health contexts has the potential to balance the traditional paternalism and power relations in health practitioner-patient relationships. The Kantian ideal endeavors to promote and protect patientsÕ autonomous decision-making Ñ to entrench competent patientsÕ rights to ultimate decisional authority.

However, in addition to inevitable problems associated with attempts to apply this kind of under-developed account, 15 there are a number of diffi- culties with appealing to this picture of Kantian autonomy in making judgments about patientsÕ capacities for autonomous decision-making.

These difficulties, which appear to stem from unanalyzed assumptions about human capacities, make this concept unfit for practical use in bioeth- ics. My main criticisms of Kantian autonomy concern: (1) its impractical- ity given the nature of patients and health contexts; (2) its normative assumptions about human nature; and (3) its potential for abuse.

First, given the nature of patienthood, the Kantian concept of autonomy demands too much of patients. This idealistic concept is of little practical relevance in health contexts where patients, on the whole, bear little re- semblance to the Kantian free, independent, exclusively rational individu- al. There are a variety of constraints common in health settings which may undermine Kantian autonomy. For example, some patients may lack or have reduced cognitive and volitional capacities which, on the Kantian view, render them imperfectly autonomous or nonautonomous (OÕNeill, 1984, p. 176). Other common internal constraints include pain, anxiety, fear, depression, effects of treatment, lack of information, inadequate un- derstanding, and false beliefs. In addition, patients generally are subject to a variety of external controlling influences. Common external constraints include pressure from relatively powerful health care professionals, and from friends and family members. Institutional environments themselves are often disorienting and restrictive, controlling patients in various ways via architecture, equipment, procedures and routines. Any one or more of BARBARA SECKER 50 such constraints may impede capacity for reasoned deliberation and choice and, therefore, Kantian autonomy.

Moreover, the highly rationalistic, individualistic Kantian account ap- pears to assume that all that patients need to qualify as autonomous, in addition to the requisite intrinsic capacities (no small feat), is negative freedom. However, patients frequently are in vulnerable positions, are unable to act on their decisions, and require that positive measures be taken on their behalf. For example, mere non-interference is insufficient for a patient who requires a professionalÕs assistance to realize her auton- omously-made choice of active euthanasia.

If we appeal to the Kantian view (based on an ideal of the self as inde- pendent and exclusively rational), very few, if any, patients will be regarded as autonomous. Actual patients are likely to be dependent or interdepend- ent, and their decision-making is not always based (exclusively) on reason.

My second concern is that Kantian autonomy appears to place a moral premium on independence. The corresponding normative assumptions about the nature of human capacities and interaction may contribute to the devaluing of those patients who may be dependent and vulnerable. As Alastair Campbell explains, if autonomy is morally valuable, and if auton- omy is equated with independence, then dependence is regarded as moral- ly inadequate and, consequently, those who are dependent are devalued (Campbell, 1991). In CampbellÕs words, the result is that:

the chronically dependent become a special problem, an embarrassment to the dominant moral value. Fragility and vulnerability, rather than being seen as appropriate parts of life from the cradle to the grave, become obstacles to be overcome by the self-sufficient man or woman.

The ÒsuccessfulÓ patient is always the one who transcends the state of patienthood (Campbell, 1991, p. 106).

The nature of patienthood, however, is partially characterized by depend- ency of one kind or another; thus, the over-valuing of autonomy as inde- pendence devalues patients who require and accept help, those who are unable to ÒtranscendÓ their patienthood. This individualistic attitude also may threaten the welfare of persons who require help but do not seek it or accept it when it is offered for fear of the stigma attached to being Ònonau- tonomous.Ó George Agich observes that this view of autonomy as inde- pendence has engendered certain ÒdefencesÓ against dependency Ñ Óa denial of need, hostility toward helpers even in the face of disabilities that require assistance from others, contempt for the real or imagined weak- ness of others,Ó and so on (Agich, 1990, pp. 12Ð13). KANTÕS CONCEPT OF AUTONOMY 51 My third criticism of the Kantian concept is that it may promote pater- nalism better than it promotes patient autonomy. 16 Because the Kantian concept is not well suited for application to patients and yet still is one of the dominant accounts, I am concerned that patients may be measured against its rigorous standards, that the majority will not Òmeasure up,Ó and that some health care professionals and institutions may attempt to justify wholesale paternalism.

Not long ago, patient autonomy was an idea foreign to health care, and physicians traditionally have held (and many still hold) that patients do not have the capacity to participate meaningfully in decision-making; thus, adopting a model of autonomy that patients have little hope of approxi- mating may play into the hands of opponents of patient autonomy. Patients who are not perceived as autonomous may not be given a choice or even be consulted about their health care options:

[The principle of respect for autonomy] does not apply to persons who are not in a position to act in a sufficiently autonomous manner Ñ perhaps because they are immature, incapacitated, ignorant, coerced, or in a position in which they can be exploited by othersÉ The behavior of nonautonomous persons may be validly controlled on grounds of benef- icence in order to protect them from harms that might result from their behavior (Beauchamp and Childress, 1989, p. 73).

This all-or-nothing conception of autonomy which the Kantian model ap- pears to embrace has the potential to deny many, if not most, patients at least some control over their health care, even when they have considera- ble capacity for agency.

In sum, the Kantian concept of autonomy does not provide adequate practical guidance for determining the kind and degree of autonomy present in patients. 17 A more applicable account of autonomy is needed for health care contexts, one which accommodates the realities of patientsÕ situa- tions; respects dependence, interdependence, and independence; and does not justify wholesale paternalism. Toward discovering a better way of understanding autonomy in health contexts, I now turn to KantÕs model of moral autonomy to assess its potential value for bioethics. BARBARA SECKER 52 VI. ON THE PRACTICAL RELEVANCE OF KANTÕS CONCEPT FOR BIOETHICS KantÕs notion of respect for the autonomy of persons has become a funda- mental principle in bioethics, albeit used to respect quite a different kind of autonomy than Kant intended. In this section, I argue that KantÕs con- cept of moral autonomy is not a concept we should attempt to Òplug intoÓ bioethicsÕ principle of respect for autonomy because of its undesirability as an ideal and its impracticability in assessment contexts.

Due to the extended emphasis on rationality Ñ to the exclusion of adequate consideration of desires and inclinations Ñ KantÕs model of autonomy is an inappropriate ideal for bioethics. (This criticism is some- what similar to my first objection to the contemporary Kantian concept of autonomy.) As Jay Katz points out, KantÕs model is built on a theoretical view of the nature of rational beings which neglects complex interrela- tions among reason, emotion and the external world (Katz, 1984, p. 110).

This inattention does not mean that Kant does not acknowledge that finite rational beings have desires and inclinations Ñ rather, it is that he judges that consideration of Òempirical psychologyÓ is not part of his task at hand. Kant writes: ÒNor need we ask on what the feeling of pleasure or displeasure rests, how desires and inclinations arise, and how, finally, maxims arise from desires and inclination under the co-operation of rea- sonÓ (FMM, p. 45).

Contrary to popular belief, Kant does not hold that a morally worthy action performed from duty is necessarily opposed to the inclinations. He does, however, think that inclinations are Òso lacking in absolute worth that the universal wish of every rational being must be indeed to free himself completely from themÓ (FMM, p. 45). As the preceding paragraph shows, Kant does not deny that maxims can be founded on inclinations; his point is that inclinations cannot be the condition of maxims if they are to be morally virtuous. Illustrative of this is the following passage from the Critique of Practical Reason:

Now it is certainly undeniable that every volition must have an object and therefore a material; but the material cannot be supposed, for this reason, to be the determining ground or condition of the maximÉ Thus the happiness of others may be the object of the will of a rational being, but if it were the determining ground of the maxim, not only would one have to presuppose that we find in the welfare of others a natural satisfaction but also one would have to find a want such as that which is occasioned in some men by a sympathetic disposition (CPrR, pp. 34Ð35). KANTÕS CONCEPT OF AUTONOMY 53 The Categorical Imperative Ñ the supreme principle of pure practical reason Ñ must be made the condition of the autonomous agentÕs maxims, and the capacity to make it so lies within the agentÕs power.

In terms of the autonomy of patients, appealing to KantÕs concept would mean that autonomous patients must act in accord with the moral law for the sake of duty. While moral autonomy is not incompatible with inclina- tions, the autonomous decision-making process is first and foremost one of pure practical reason; according to Kant, decisions and actions based on inclinations rather than on reason are heteronomous. KantÕs idealized au- tonomous agent, successfully checking desires and dispositions against the law of pure practical reason, clearly is out of place among patients who often are physically or psychologically vulnerable, and who may have impaired cognitive and volitional capacities. Patients have little hope of approximating such an ideal Ñ their abilities fall short of the enhanced capacities of KantÕs autonomous agent.

It is doubtful that Kant actually expects finite rational beings to achieve this ideal, but he did think it a vision worthy of emulation. 18 I am not convinced, however, that we ought to counsel patients to strive in this manner. The capacity for reason is only one facet of human nature: we are more than our rational faculties. In addition, to divide our faculties into reason and inclination seems artificial, even misguided, since, as Susan Wendell expresses it, they are Òso intermingled and integrated in most human activities that such a division is rarely applicable and usually ob- scures the complexity of real people and their behaviourÓ(Wendell, 1987, p. 79). Moreover, valuing reason over inclination seems wrong-headed, since both are equally important for our survival and the quality of our human experiences. A fuller conception of patient autonomy is desirable for bioethics, one that more accurately reflects human experienceÑone which is more synergistic, as opposed to normatively dualistic, in its treat- ment of our inclinations and capacity for reason.

Furthermore, we ought to consider whether an ideal of moral autonomy is desirable in bioethics. Ought health practitioners 19 to attempt to assess or monitor the moral autonomy of their patients in an effort to respect patient autonomy and promote patient welfare? Is it appropriate for health practi- tioners to judge patientsÕ moral values, choices or actions? In my view this would be an inappropriate goal. Health practitioners do not have any special qualifications for the role of moral expert or Òmoral police.Ó Even if it were possible to prepare them for such a role, the paternalism this model would make way for is inimical to KantÕs duty-based philosophy of freedom. 20 Assuming some remain unconvinced that KantÕs concept is not a wor- thy ideal, his concept certainly would need supplementation. Contempo- BARBARA SECKER 54 rary concern about respecting patient autonomy is motivated primarily by concern to promote and protect the rights and well-being of patients. While a model of moral autonomy like KantÕs may promote patientsÕ moral well- being, it may fall short in promoting physical, psychological or social well-being. In addition, the right to make morally autonomous decisions (that is, to recognize the validity of the moral law and decide to accept it for the sake of duty) is a rather narrow right considering the range of decision-making patients are confronted with in health contexts. Some decisions patients face arguably are moral decisions Ñ for example, wheth- er to have an abortion, request physician-assisted suicide, or to be truthful about medical history or particular diagnoses. However, many decisions would appear to be a matter of personal preference rather than a matter of morality Ñ for instance, whether to chose drug A or drug B, drug A or surgery, lumpectomy or mastectomy, treatment or further observation, to remain at home or go to a nursing home. 21 For most patients, the primary goal in interacting with health practitioners is not to deliberate about or revise their moral values, but rather to make decisions about and receive health care (Emanuel and Emanuel, 1995, p. 74).

For the reasons given above, then, KantÕs concept of moral autonomy would appear, at worst, altogether inappropriate and, at best, inadequate.

For those who are not persuaded, I turn now to consider whether KantÕs concept of moral autonomy would be amenable to assessment in health contexts. Could we appeal to KantÕs concept in order to determine which patients have the capacity for moral autonomy and/or which patients would be deciding or acting in a morally autonomous way with regard to particu- lar health-related decisions?

In KantÕs view, finite rational beings have the capacity for moral auton- omy; this capacity is an inherent, general feature of such beings, but may or may not be exercised via particular decisions or actions. It is not clear whether Kant would consider patients rational beings (given their vulner- abilities and often diminished capacities), but it is likely that certain groups of patients Ñ infants, young children, those who are severely demented or comatose, for example Ñ would not be counted by Kant as rational be- ings. Assuming Kant would consider rational the remainder of the class of patients, we can continue to consider the applicability of his model for these patients at least.

If we were able to detect KantÕs distinct brand of rationality in patients, then we would know they have the capacity for moral autonomy. Since, for Kant, to be a rational being is to have the capacity for autonomy, we can simplify this discussion by setting out KantÕs criterion for the latter:

beings have the capacity for autonomy if and only if they have the capacity KANTÕS CONCEPT OF AUTONOMY 55 to act on self-given, yet universal, principles (even when inclination is absent). Autonomous agents are able to take the Categorical Imperative as their fundamental maxim.

Can we use KantÕs Categorical Imperative, then, to classify the maxims patients adopt in order to determine whether they have the capacity for autonomy and/or whether in a particular situation they would be deciding/ acting in a morally autonomous way? The logistics of attempting to use KantÕs concept to assess the moral autonomy of patients Ñ which we do in an effort to determine whether to act, first and foremost, to respect patient autonomy or to protect patient well-being Ñ present some interrelated epistemic and practical problems. We do not have direct access to pa- tientsÕ maxims, and neither can patients themselves, according to Kant, be absolutely certain about the nature and moral status of their maxims. The Òopacity of the human heartÓ undermines any attempt to assess the univer- salizability of maxims. Given that maxims are not observable, all we would be able to do is to evaluate what patients tell us about their maxims and perhaps how they act, 22 and attempt to infer whether their maxims con- form to the Categorical Imperative. But, maxims are not reliably inferable given that patients have imperfect self-knowledge, that the deception of self and others is possible, and that practitioners may misinterpret what they see and hear. In OÕNeillÕs words, KantÕs universalizability test is Òprimarily of use to agents in guiding their own moral deliberations, and can only be used most tentatively in assessing the moral worth of othersÕ action, where we are often sure only about specific outward aspects of action and not about the maximÓ (OÕNeill, 1989, p. 103).

The problem of inference, however, is not unique to Kant (although we would need to wrestle with the problem in a unique way if we were to adopt his model). These limits to knowledge of self and others constitute the context in which we must make practical judgments about patientsÕ moral autonomy. The best we would be able to do, then, would be to assess patients by the maxims they reveal as their own in light of our ideal of the maxims of fully autonomous agents. If patientsÕ maxims appeared to pass the universalizability test, we would have to assume, for practical purpos- es, that conceptual and volitional inconsistencies had been ruled out, and thus that their maxims were at least not morally unacceptable. However, we could not know whether they would be autonomous in choosing or acting on that maxim for we cannot know whether they acted from duty or merely in accord with duty.

A potential solution to this problem might be to assume that the class of mature, conscious, non-demented patients has the capacity for moral au- tonomy, and respect the decisions and actions of such patients. However, BARBARA SECKER 56 if we are concerned that certain features of patienthood and health con- texts may interfere with patients deciding or acting in morally autonomous ways in general or in particular instances, this assumption will not suit our purposes. We would want to know whether patients have the capacity to act autonomously in a particular situation or whether they would be acting autonomously, but KantÕs concept of moral autonomy lies beyond the realm of empirical observation. We cannot assess the moral autonomy of patients because moral autonomy is transempirical; in other words, we cannot know anything, strictly speaking, about a personÕs moral autono- my.

If we are committed to expressing respect for the dignity of patients as morally autonomous beings, given the transempirical nature of KantÕs concept we must strive to do this by recognizing patient self-determina- tion, specifically decision-making capacity, which is more readily subject to empirical assessment. Kant would approve of this conclusion since, for him, there is an important connection between moral autonomy and self- determination. Respecting a personÕs capacity for self-determination is a way of expressing respect for her dignity as a morally autonomous being.

The principle of respect for self-determination, then, should be used to bring about an empirical order of right, one which points indirectly to the dimension of moral autonomy which lies outside the empirical realm.

The lack of practical relevance of KantÕs concept of moral autonomy 23 may well have led contemporary Kantian bioethicists to substitute the so- called Kantian concept of autonomy for KantÕs original. While the transem- pirical nature of KantÕs concept necessitates the move to respect patient self-determination, we need not appeal to the Kantian notion of autonomy as self-determination, outlined and critiqued above.

VII. PROLEGOMENA TO A NEW PARADIGM So far, I have disentangled KantÕs original concept of moral autonomy from the contemporary Kantian concept of autonomy found in bioethics. I also have argued that neither concept ought to be appealed to in efforts to respect the autonomy of patients. In this final section, I hope to advance bioethics discourse by suggesting a direction we might pursue towards reconceptualizing patient autonomy.

The general direction I suggest is characterized by a contextual, interac- tive understanding of autonomy Ñ one in which autonomy is not equated with independence, deliberative rationality or negative liberty. Beauchamp (1991), however, appears to hold that certain criteria Ñ much like those of KANTÕS CONCEPT OF AUTONOMY 57 the Kantian concept described earlier Ñ are central to the notion of auton- omy:

Whatever may in addition be packed into the concept, the autonomous person is one who is capable of independent, intentional, informed, and reasoned judgments and actions. Any viable theory of autonomous per- sons must accept something like these properties as necessary condi- tions of autonomy, because of their centrality to the concept (Beau- champ 1991, p. 60).

It is not at all clear to me, however, that such features are necessary conditions of autonomy. While different accounts of autonomy share an abstract conceptual core Ñ the idea of self-government or self-determina- tion Ñ suggested by the etymology of the term, the content or criteria of a concept of autonomy will depend on the underlying view of the self.

For the reasons given above, a defensible concept of patient autonomy would need to be built on a conception of the self which is significantly different from the rational individualist view of the self underlying the Kantian concept of autonomy. As Sherwin states (and as I argued earlier), Òactual people are not independent, and their decision-making does not always meet the norms that define rationalityÓ (Sherwin, 1992, p. 137).

The alternative concept of patient autonomy I have in mind appeals to a more realistic view of the self as both separate and connected Ñ as neither wholly socially constituted, nor completely independent. This self is de- fined, in part, by its socio-political context and its relationships with oth- ers. On this view, people often make choices in the company (even with the assistance) of others who help shape their lives Ñ their characters, beliefs, values and goals (see also Sherwin, 1992, pp. 53, 137Ð157; Hoag- land, 1988, pp. 144Ð146).

HoaglandÕs description of a self-in-community may be useful in articulat- ing some of the features of a view of the self which would give rise to a better model of autonomy with practical relevance for bioethics. 25 In con- structing her account of moral agency and choice, Hoagland invokes Òa self who is both elemental and related, who has a sense of herself making choic- es within a context created by communityÓ (Hoagland 1988, p.145). For Hoagland, the self is not essentially defined in terms of others, but it does emerge though interactions with them. HoaglandÕs term for this separate yet connected self is ÒautokoenonyÓ which means Òthe self in communityÓ: 26 An autokoenonous being is one who is aware of herself as one among others within a community that forms her ground of being, one who BARBARA SECKER 58 makes her decisions in consideration of her limitations as well as in consideration of the agendas and perceptions of others. She does not merge with others, nor does she estrange herself; she interacts with others in situations (Hoagland, 1988, p. 145; her emphasis).

Sherwin describes what HoaglandÕs ÒautokoenonyÓ might mean in health contexts:

autokoenony suggests an understanding of patients existing in a social world, where their ends and activities are defined in conjunction with others they trust. This view offers a more realistic perspective of pa- tients choosing in the company of others who help shape their lives; when patients are confronted with difficult decisions, physicians and patients might include in the decision-making process those others who are trusted by the patient (Sherwin, 1992, p. 156). 27 An alternative account of autonomy based on a view of self-in-communi- ty, as understood by thinkers like Hoagland and Sherwin, 28 would appear to have practical relevance for actual patients and to avoid some of the problems associated with the concepts of Kant and contemporary Kan- tians. The alternative model I envision would not require that patients have freedom of action Ñ that is, the ability to make their decisions effective in action. It would separate the question of whether patients are capable of autonomous decisionmaking from the question of whether they are able to effectuate their decisions, and it would focus on the former. As Hoagland argues, that we are unable to control certain factors which affect us, and which may determine whether we can carry out our decisions, does not mean that we are unable to make choices. An adequate concept of autonomy must focus on our ability to make decisions:

choice is at the very core of the concept of Ômoral agency.Õ It is not because we are free and moral agents that we are able to make moral choices. Rather, it is because we make choices, choose from among alternatives, act in the face of limits, that we declare ourselves to be moral beings. That is what it means to be a moral being....[M]aking choices within limited situations is a matter of affirming moral agency, not undermining it (Hoagland, 1988, pp. 230Ð231; her emphasis).

Such a model of decisional autonomy would not demand independence or self-sufficiency but rather would be compatible with certain dependencies and interdependencies. Agich brings to light some important interrelation- ships among autonomy, dependence and interdependence: KANTÕS CONCEPT OF AUTONOMY 59 Maintaining a sense of autonomous well-being is consistent with de- pendencies on medication or professional care if those dependencies help to maintain a sense of functional integrity in the areas of life that individuals value (Agich, 1990, p. 16).

Agich (1990) cites an example of a person who uses a wheelchair and who requires assistance from others in a number of activities of daily living.

This assistance, however, enables her to work for a particular charitable organization with which she identifies strongly. Her dependency does not interfere with (in fact, it promotes) her participation in what has meaning to her, and her inability to undertake certain tasks alone Ñ such as shop- ping or bathing Ñ is less significant to her than her ability to work. Thus, certain dependencies and interdependencies in personal and professional relationships can be empowering Ñ they actually can enhance meaningful autonomy.

This alternative concept of decisional autonomy Ñ given the view of the self-in-community on which it is based Ñ would characterize deci- sionmaking as cooperative rather than antagonistic. This would help to undermine the present struggle for control over decisionmaking. That is, the main issue would no longer be whether patients ought to make the decision (autonomy) or whether the decision ought to be given to the health practitioner (beneficence or paternalism) but, rather, how to empower patients to make their decisions, how to foster decisional autonomy. 29 The main concern, then, would not be whether patients are autonomous or nonautonomous Ñ whether they are to be allowed to par- ticipate or not in decisions about their health care Ñ but, rather, whether all that could be done to enhance autonomous decisionmaking capacity has been done and whether patients can use these resources to make a decision. 30 The emphasis on enhancing the autonomous decisionmaking capacity of patients would bring a positive dimension to the concept of autonomy (lacking at least in the concept of contemporary Kantians). That is, if we really are committed to fostering autonomous decisionmaking capacity and its realization, then, in many (if not most) cases, more is required than the negative dimension of noninterference. This idea is captured in Ag- ichÕs call for a richer concept of autonomy which would view Òchoice as a problem of providing options that are meaningful rather than as an issue of removing obstacles to choice or impediments to actionÓ (Agich, 1990, p.

12). An adequate alternative approach to autonomy in health care settings would demand the provision of resources needed for the exercise of pa- tientsÕ autonomously-made decisions. This would involve looking beyond BARBARA SECKER 60 the artificial boundaries of Òthe health care contextÓ to the larger moral and socio-political context.

The set of the criteria I outline below is my attempt to capture the general nature of the alternative model of autonomy sketched above. On this contextual, interactive view of decisional autonomy, a patient who chooses autonomously is able to:

1. recognize the nature of the decision to be made (i.e., that there are options and what those options are), alone or with the assistance of others; 2. respond to the need to make a decision by undertaking to make such choice, alone or with the assistance of others; 3. communicate and interact with others, if necessary, toward making a decision; 4. understand information relevant to the decision to be made and use this understanding to consider the likely pros and cons of available options, alone or with the assistance of others; 5. arrive at a decision, alone or with the assistance of others, which (a) is free from central contradiction, and (b) follows an uncoerced decision- making process that avoids unresolvable means-ends errors in reason- ing; and 6. resolve to make the decision effective in action alone or with the assist- ance of others.

To summarize, this alternative account of autonomy can be characterized as 1) procedural (versus substantive), 2) decisional (rather then execution- al), 3) contextual and interactive (versus abstract and individualistic), and 4) decision-specific (as opposed to global or general). To elaborate briefly on this characterization, this procedural model recognizes more than one conception of Òthe good life;Ó it does not necessarily preclude choosing any particular end in recognizing decisional autonomy. Its procedural re- quirements are less constraining (than with either KantÕs or the Kantian concept), reflecting a broader understanding of autonomous decision-mak- ing and allowing changes in patientsÕ beliefs, values and goals. It does not equate autonomy with morality or with deliberative rationality. As long as patients are able to understand and consider information relevant to the decision at hand, they may base their decision on this understanding and consideration, or on other considerations, such as emotion, inclination, moral or other values, religious beliefs, and/or the opinions of others who may be involved in the decision-making process.

In focusing on decisional rather than executional autonomy, this con- cept does not require freedom of action and may be better able (than more KANTÕS CONCEPT OF AUTONOMY 61 robust models) to help identify social and cultural factors affecting pa- tientsÕ abilities to exercise their autonomously-made decisions. Decisional autonomy is, in part, a relational, not simply an intrinsic, property of patients. Further, on this contextual, interactive model, dependence and interdependence do not necessarily preclude autonomous decision-mak- ing; in fact, certain relationships so characterized may even be necessary for such decision-making. Finally, the label ÒautonomousÓ is affixed to particular instances of patientsÕ choosing Ñ autonomy is decision-specif- ic, rather than global. That is, it is understood as a feature of patients in some aspects of their life, rather than as a feature of whole persons over their entire life.

Questions this alternative model must address, in addition to its theoret- ical underdevelopment, include: the potential for conflict, manipulation or unwarranted paternalism (given the emphasis on cooperative decision- making); the role of socialization; the problem of inferring understanding (refer to criterion 4); and the designation of a threshold for decisional autonomy. In addition, as Sherwin points out, a model of autonomy that focuses on empowering patients Ñ on strengthening their agency Ñ to make decisions that benefit them will Òrequire radical rethinking of the patient-physician relationship and development of improved patterns of communication and mutual respectÓ (Sherwin, 1992, p. 156).

In conclusion, the principle of respect for the autonomy of personsÑa legacy of KantÑis of fundamental importance in bioethics. However, we need a concept that is adequate for assessing patient autonomy as deci- sion-making capacity in health contexts. Reconceptualization of this key conceptÑmediated by a commitment to practical relevanceÑmust be un- dertaken toward providing bioethics with a workable, defensible model of patient autonomy. I submit that it would be fruitful to begin along the lines sketched above.

ACKNOWLEDGEMENTS I would like to thank Cheryl Cline, Heiner Bielefeldt, William Harvey, Jan Marta, Kath- ryn Morgan, Gordon Nagel, Bjorn Ramberg, Michael Stingl, Susan Wendell and three anonymous reviewers for their comments on earlier drafts or sections of this paper. I am grateful to the Social Sciences and Humanities Research Council of Canada for its support from 1994Ð1997 during which time I produced this manuscript. BARBARA SECKER 62 NOTES 1. As Tom Beauchamp points out, the notion of autonomy has become so diverse that it is often used to refer to a duty, a right, a freedom, a disposition, or an action. He identifies the following explications of the central meaning of ÒautonomyÓ in con- temporary moral philosophy: 1) ÒauthenticityÓ; 2) Òobedience to self-prescribed lawÓ; 3) Òobedience to moral lawÓ; 4) Òpersonal choiceÓ; 5) Òthe freedom to chooseÓ; 6) Òhaving preferences about oneÕs preferencesÓ; 7) Òchoosing and creating oneÕs own moral positionÓ; 8) Òmental healthÓ; 9) ÒconscientiousnessÓ; 10) Òresponsible ac- tionÓ; and 11) Òaccepting responsibility for oneÕs views and actionsÓ (Beauchamp, 1991, p. 63).

2. For example, a BIOETHICSLINE database search (i.e., a computerized literature retrieval system) for Òautonomy or informed consent or competenceÓ paired with ÒKant(ian)Ó reveals three times as many citations than when paired with ÒJohn Stuart MillÓ and twice as many than when paired with ÒGerald Dworkin.Ó For some exam- ples of bioethics books and articles which engage with a Kantian concept of autono- my, refer to Mappes and DeGrazia (1996), and the references listed in note 10.

3. Beauchamp and Childress, however, write: ÒIn contemporary biomedical ethics the word autonomy typically refers to what makes a life oneÕs own; viz. that it is shaped by personal preferences and choices. This conception of autonomy is emphatically not KantÕsÓ (Beauchamp and Childress, 1996, p. 58). The authors then provide a brief two-paragraph description of KantÕs contrast between moral autonomy and heterono- my. However, what the authors describe as autonomyÕs typical referent in bioethics is not the contemporary Kantian conception of autonomy (which I describe in the sec- tion titled ÒThe Kantian Concept of AutonomyÓ). Because it is more common to conflate KantÕs concept of autonomy and the contemporary Kantian one, I focus on making this distinction and analyzing in depth its relevance to the principle of respect for autonomy in bioethics.

4. I am indebted to OÕNeill (1989; 1991), Hill (1989), Wolff (1973), Beck (1960) and Allision (1990; 1996) whose commentaries have helped structure my thinking about KantÕs moral philosophy.

5. In this section, I provide a brief account of what I take to be KantÕs concept of autonomy. I am aware that there is considerable controversy over how he conceived autonomy, and that certain of his claims about autonomy are often questioned by critics. Due to limitations of space, this section is more explanatory than critical; I reserve my critical discussion for the next section since the purpose of this paper is to assess the practical relevance of KantÕs concept of autonomy for bioethics.

6. To avoid confusion, I will refer to the primary works of Kant by name (abbreviated), rather than by author and date: Foundations of the Metaphysics of Morals (ÒFMMÓ); Critique of Practical Reason (ÒCPrRÓ); The Doctrine of Virtue (ÒDVÓ).

7. Later, in the Critique of Practical Reason, Kant argues against the possibility of such a derivation, and claims instead that autonomy, freedom and the categorical impera- tive presuppose or imply each other.

8. One acts heteronomously if one fails to act for the sake of duty. If such an act is in accord with the moral law, oneÕs action is morally correct but without moral worth; if the act does not conform to the moral law, it is morally reprehensible.

9. I am interested in Kantian autonomy as presented in influential bioethics textbooks because it is in these volumes that at least some space is devoted to explaining ethical concepts, principles, and theories; also, these are the theoretical tools that the most visible academic bioethicists are handing to clinicians and students of bioethics. I KANTÕS CONCEPT OF AUTONOMY 63 selected this passage because it gives a relatively detailed description of so-called Kantian autonomy. Sketchier accounts bearing the label ÒKantianÓ can be found in, e.g., Beauchamp and Walters (1994, p. 23) and Munson (1996, p. 40).

10. For some examples which appear similar to the Mappes and DeGrazia account, refer to: Gauthier (1993, pp. 23Ð24); Jinnett-Sack (1993, pp. 98Ð103); Munson (1996, pp.

40, 59); Nash (1990, p. 118); and Norden (1995, pp. 180Ð182).

11. GauthierÕs description is similar: ÒWe recognize that as a free and rational being the other [person] has the capacity to choose his or her own goals and projects on the basis of moral principles known by reason and, thus, to act on a personal conception of what is right. Only when we respect and do not interfere with othersÕ goals, projects, and actions, chosen by their own conception of what is right, are we respect- ing their autonomy as rational agentsÓ (1993, p. 24; my emphasis).

12. Jinnett-Sack states that, in bioethics, the predominant view of the autonomous sub- ject is of one who is Òsolitary, imminently educable and capable of self-knowledge;Ó she cites Iris Murdoch as attributing this view to the ÒmarriageÓ of Kant, utilitarian- ism and modern psychology (1993, p. 101; my emphasis). Moreover, Mappes and DeGraziaÕs description of the autonomous individual as able to resist manipulation or coercion by others suggests a considerable degree of independence (1996, p. 28).

13. See also Herman, 1991, p. 792.

14. What accounts for the confusion or gap between KantÕs notion of autonomy and that of contemporary Kantians writing in bioethics? OÕNeill suggests that this is partly due to the fact that Òmodern protagonists of ÔKantianÕ ethics are mainly, some of them exclusively, interested in rights,Ó whereas rights are only one element of KantÕs duty-based philosophy of freedom (OÕNeill, 1989, p. 66).

15. My focus here is the applicability and adequacy of the Kantian concept in health care contexts. I am more interested in the practical problems involved in applying this concept than I am in its more abstract conceptual flaws; thus, I will be concerned with the latter only to the extent that they affect or create the former.

16. This criticism is not without basis. Refer, for example, to KomradÕs (1983) ÒA de- fence of medical paternalism: Maximizing patientsÕ autonomy.Ó Also refer to Cassell (1977, p. 17) who holds that illness impairs a personÕs authentic self, independence, knowledge, understanding, and reason, thereby significantly diminishing autonomy.

He proposes that Òthe best way to preserve autonomy is to cure the patient of the disease that impairs autonomy and return him to his normal lifeÓ which may well mean treating the patient against his will (Cassell, 1977, p. 18).

17. For further development of the above criticisms and other critiques of Kantian auton- omy (or similar accounts), refer to: Agich (1990); Hoagland (1988); Kittay and Mey- ers (eds.) (1987, ch. 8 and 9); Sherwin (1992, ch. 7); and Weinberg (1988).

18. The question whether Kant meant for his concept of autonomy to be normative for the moral life of human beings is somewhat beside the point Ñ the purpose of having a concept of autonomy in bioethics is a normative one.

19. For the purposes of this argument, I include bioethicists in the class of health practi- tioners.

20. For a discussion of the potential promise and pitfalls of a similar model Ñ the deliberative model of physician-patient relationship Ñ refer to Emanuel and Emanuel (1995). In this model, Ò[t]he physicianÕs objectives include suggesting why certain health-related values are more worthy and should be aspired toÉ.The conception of patient autonomy is moral self-development; the patient is empowered not simply to follow unexamined preferences or examined values, but to consider, through dia- logue, alternative health-related values, their worthiness, and their implications for BARBARA SECKER 64 treatmentÓ (Emanuel and Emanuel, 1995, p. 69). While the focus here is on worthi- ness of health-related values rather than the worthiness of maxims in relation to the Categorical Imperative, both models risk paternalism since attempts to persuade may well be experienced by patients as attempts to manipulate or coerce, given power dynamics in favor of the professional in the practitioner-patient relationship.

21. Another way of expressing this might be to say that many health care decisions would appear to revolve around hypothetical rather than categorical imperatives.

22. I say ÒperhapsÓ because inferring maxims from actions is not foolproof. For example, a patient might comply with her physicianÕs ÒordersÓ for the sake of the duties of self- preservation and promise-keeping; whereas, another patient might comply with her physicianÕs orders because she is simply in the habit of following the orders of authority figures or because she fears the wrath of her doctor or family members.

Both patients perform the same actions of compliance, but, according to Kant, the first patient who acts from duty acts autonomously while the second patient who acts only in accord with duty acts heteronomously.

23. To say that appealing to KantÕs model of autonomy is undesirable and impracticable with respect to patients, is not to say that KantÕs ideal of the morally autonomous agent has no place whatsoever. In fact, KantÕs model might be of value for health care professionalsÑwho generally are in less vulnerable or compromised positions than patientsÑand who frequently find themselves in moral quandaries regarding their duties as professionals. However, given that Kant provided little guidance on how morally relevant maxims are to be formulatedÑand given that different formulations may yield different results when tested for consistencyÑit is not clear that KantÕs Categorical Imperative can help us decide whether an action is morally required, morally forbidden, or morally permissible. For further discussion of this problem refer, for example, to Nell (1975), OÕNeill (1989), Wolff (1973), and Hoffmaster (1990, pp. 245-46).

24. Given the context of this passage, it seems clear that Sherwin is referring to tradition- al philosophical notions of rationality Ñ instrumental rationality and rationality of ends.

25. Naturally, a more developed account of the nature of the self in community is needed both to Òfill outÓ and to justify such an alternative model. Central questions include:

In what ways, and to what extent, is self-identity constituted or defined by communi- ty? In what ways, and to what extent, is it chosen? How do social experience and personal choice interact? What, if any, moral consequences follow from responses to these questions? (Pursuing any one of these questions leads one into issues of great complexity, well beyond the scope of this paper.) 26. Greek, ÒautoÓ (self), ÒkoinoniaÓ (community, or any group whose members have something in common) (Hoagland, 1988, p. 145).

27. Since not all patients have people who care about them, that they can trust and rely on, it is important that Sherwin adds: ÒPatients who are isolated, that is, who have no others they can trust, could be helped to form relationships that would foster their decision-making in an interactive way. Self-help groups of patients with a common condition, for instance, usually provide patients with the opportunity to explore the complexities of their decisions in a nonhierarchical environmentÓ (Sherwin, 1992, p.

156).

28. For detailed development of an alternative view of autonomy as relationalÑbuilt on a relational view of the self Ñ see Sherwin, 1998.

29. This is not to say that reconceiving autonomy along these alternative lines would itself eliminate the conflict between patients and practitioners over decisionmaking KANTÕS CONCEPT OF AUTONOMY 65 power. An alternative model of autonomy which assigns less theoretical weight to such conflict may contribute to diffusing this Òpower struggle,Ó but, clearly, before the focus in health care moves from awarding decisional authority to enhancing autonomous decisionmaking capacity, certain compatible political, economic and educational structures would need to be in place.

30. Note that it does not follow that the patient who is unable to make a decision even after every effort has been made to foster his/her decisionmaking capacity should be excluded altogether from the decisionmaking process. Patients should be encouraged to participate in decisions about their health care to the fullest extent that they are able. REFERENCES Agich, G.J. (1990). ÔReassessing autonomy in long-term care,Õ Hastings Center Report (November/December), 12Ð17.

Allision, H. (1990). KantÕs Theory of Freedom, Cambridge University Press, Cambridge, England.

Allision, H. (1996). Idealism and Freedom: Essays on KantÕs Theoretical and Practical Philosophy, Cambridge University Press, Cambridge, England.

Beauchamp, T.L. (1991). ÔCompetence,Õ in M.G. Cutter and E.E. Shelp (eds.), Compe- tency: A Study of Informal Competency Determinations in Primary Care, Kluwer Academic Publishers, Dordrecht, pp. 49Ð77.

Beauchamp, T. and Childress, J. (1989). Principles of Biomedical Ethics (3rd edition), Oxford University Press, New York.

Beauchamp, T. and Childress, J. (1994). Principles of Biomedical Ethics (4th edition), Oxford University Press, New York.

Beauchamp, T. and Walters, L. (eds.) (1994). Contemporary Issues in Bioethics (4th edi- tion), Wadsworth Publishing Company, Belmont, California.

Beck, L.W. (1960). A Commentary on KantÕs Critique of Practical Reason, The University of Chicago Press, Chicago.

Campbell, A. (1991). ÔDependency revisited: The limits of autonomy in medical ethics,Õ in M. Brazier and M. Lobjoit (eds.), Protecting the Vulnerable: Autonomy and Con- sent in Health Care, Routledge, London and New York, pp. 101Ð112.

Cassell, E.J. (1977). ÔThe function of medicine,Õ Hastings Center Report (December), 16Ð 19.

Christman, J. (ed.) (1989). The Inner Citadel, Oxford University Press, New York.

Emanuel, Ezekiel J. and Emanuel, Linda (1995). ÔFour models of the physician-patient relationship,Õ in John D. Arras and Bonnie Steinbock (eds.), Ethical Issues in Modern Medicine, Mayfield Publishing Company, Mountain View, California, pp.

67Ð77.

Gauthier, Candace Cummins (1993). ÔUnderstanding the philosophical foundations of the principle of respect for autonomy,Õ Kennedy Institute of Ethics Journal 3(1), 21Ð 37.

Gillon, R. (1985). Philosophical Medical Ethics, Wiley, London.

Herman, B. (1991). ÔAgency, attachment, and difference,Õ Ethics 101, 775Ð797.

Hill, T.E. Jr. (1989). ÔThe Kantian conception of autonomy,Õ in J. Christman (ed.), The Inner Citadel, Oxford University Press, New York, pp. 91Ð105.

Hoagland, S. (1988). Lesbian Ethics: Toward New Values, Institute of Lesbian Studies, Palo Alto, California. BARBARA SECKER 66 Hoffmaster, B. (1990). ÔMorality and the social sciences,Õ in George Weisz (ed.), Social Science Perspectives on Medical Ethics, Kluwer Academic Publishers, Dordrecht, pp. 241Ð260.

Jinnett-Sack, S. (1993). ÔAutonomy in the company of others,Õ in Andrew Grub (ed.), Choices and Decisions in Health Care, John Wiley & Sons, New York, pp. 97Ð136.

Kant, I. (1959). Foundations of the Metaphysics of Morals, L.W. Beck (trans.), The Bobbs- Merrill Company, Inc., Indianapolis.

Kant, I. (1964). The Doctrine of Virtue: Part II of the Metaphysic of Morals, M.J. Gregor (trans.), University of Philadelphia Press, Philadelphia.

Kant, I. (1993). Critique of Practical Reason (3rd edition), L.W. Beck (trans.), MacMillan Publishing Company, New York.

Katz, J. (1984). The Silent World of Doctor and Patient, The Free Press, New York.

Kittay, E. Feder and Meyers, D. (eds.) (1997). Women and Moral Theory, Rowman and Littlefield, Totowa, N.J.

Komrad, M.S. (1983). ÔA defence of medical paternalism: Maximizing patientsÕ autonomy,Õ Journal of Medical Ethics 9, 38Ð44.

Mappes, T.A. and DeGrazia, D. (eds.) (1996). Biomedical Ethics (4th edition), McGraw- Hill, Inc., New York.

Morreim, E.H. (1991). ÔCompetence: At the intersection of law, medicine, and philoso- phy,Õ in M.G. Cutter and E.E. Shelp (eds.), Competency: A Study of Informal Competency Determinations in Primary Care, Kluwer Academic Publishers, Dordrecht, pp. 93-125.

Munson. R. (ed.) (1996). Intervention and Reflection: Basic Issues in Medical Ethics (5th edition), Wadsworth Publishing Co., Belmont, California.

Nell, O. (1975). Acting On Principle, Columbia University Press, New York.

Nash, P. (1990). ÔAutonomy, competence and mental disorders,Õ in Donald Evans (ed.), Why Should We Care? St. MartinÕs Press, New York, pp. 116Ð127.

Norden, M. (1995). ÔWhose life it is anyways? A study in respect for autonomy,Õ Journal of Medical Ethics 21, 179Ð183.

OÕNeill, O. (1984). ÔPaternalism and partial autonomy,Õ Journal of Medical Ethics 10, 175Ð80.

OÕNeill, O. (1989). Constructions of Reason: Explorations of KantÕs Practical Philosophy, Cambridge University Press, New York.

OÕNeill, O. (1991). ÔKantian ethics,Õ in P. Singer (ed.), A Companion to Ethics, Blackwell Publishers, Oxford, pp. 175-185.

Sherwin, S. (1992). No Longer Patient: Feminist Ethics and Health Care, Temple Univer- sity Press, Philadelphia.

Sherwin, S. (ed.) (1998). The Politics of WomenÕs Health: Exploring Agency and Au- tonomy, Temple University Press, Philadelphia.

Weinberg, J.K. (1988). ÔAutonomy as a different voice: Women, disabilities and deci- sions,Õ in M. Fine and A. Asch (eds.), Women with Disabilities: Essays in Psychol- ogy, Culture, and Politics, Temple University Press, Philadelphia, pp. 269Ð296.

Wendell, S. (1987). ÔA (qualified) defense of liberal feminism,Õ Hypatia 2, 65Ð93.

Wolff, R.P. (1973). The Autonomy of Reason: A Commentary on KantÕs Groundwork of the Metaphysic of Morals, Harper & Row Publishers, New York. Copyright ofJournal ofMedicine &Philosophy isthe property ofOxford University Press/ USA anditscontent maynotbecopied oremailed tomultiple sitesorposted toalistserv without thecopyright holder'sexpresswrittenpermission. However,usersmayprint, download, oremail articles forindividual use.