Journal of Applied Philosophy,Suicide Vol. 19,Intervention No. 3, 2002 and Non-Ideal Kantian Theory
245
Suicide Intervention and Non-Ideal Kantian Theory
MICHAEL J. CHOLBI
Philosophical discussions of the morality of suicide have tended to focus on its justifiability from an agent’s point of view rather than on the justifiability of attempts by others to intervene so as to prevent it. This paper addresses questions of suicide intervention within a broadly Kantian perspective. In such a perspective, a chief task is to determine the motives underlying most suicidal behaviour. Kant wrongly characterizes this motive as one of self-love or the pursuit of happiness. Psychiatric and scientific evidence suggests that suicide is instead motivated by a nihilistic disenchantment with the possibility of happiness which, at its apex, results in the loss of the individual’s conception of her practical identity. Because of this, methods of intervention that appeal to agents’ happiness, while morally benign, will prove ineffective in forestalling suicide. At the same time, more aggressive methods violate the Kantian concern for autonomy. This apparent dilemma can be resolved by seeing suicide intervention as an action undertaken in non-ideal circumstances, where otherwise unjustified manipulation, coercion, or paternalism are morally permitted.
1. The chances are good that each of us will some day confront a person close to us contemplating suicide. Every year in the United States, suicide attempts lead to 30,000 deaths and nearly half a million visits to emergency rooms [1]. This number does not even include another still larger group that contemplates suicide or forms suicidal intentions but never actually initiates a suicide attempt. And for each such suicidal person, there are numerous other persons — friends, family members, and health care providers among them — for whom a troubling moral question should arise: How, if at all, may I act in order to prevent another person from taking her own life? By and large, traditional philosophical discussions of the morality of suicide generally neglect this question in favor of another: Is an agent ever morally justified in committing suicide? [2] In so doing, philosophers have tended to pay scant attention to the psychological complexities associated with suicidal behaviour. Studies have shown repeatedly that suicidal thoughts and attempts are strongly associated with depression, schizophrenia, or similar mental illnesses [3]. Yet philosophers have often misunderstood the psychology of most suicides by blithely assuming that suicidal persons are psychologically typical. Furthermore, the psychological aspect of suicide is vital to our understanding of the morality of suicide intervention. That suicidal persons are not obviously functioning normally will presumably affect not only when intervention is morally permissible but also how such intervention ought to be undertaken. The aim of my discussion is to identify some of the problems that the psychological features of suicidal persons might raise for a Kantian answer to the question, what © Society for for Applied Applied Philosophy, Philosophy,2002, 2002 Blackwell Publishers, 108 Cowley Road, Oxford, OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA.
JOA19.3C05
245
11/7/02, 15:47
246
Michael J. Cholbi
means of intervention to prevent suicide are morally permissible? (I shall use the notion of ‘intervention’ broadly, to encompass any act undertaken to prevent or delay another from taking her own life.) I will argue that certain psychological features strongly associated with suicidal behaviour may sometimes necessitate methods of intervention that violate suicidal individuals’ autonomy in ways Kantians generally find objectionable. But this conclusion can be avoided by seeing the threat of suicide as a non-ideal circumstance to which Kantian moral theory can be applied. In particular, once armed with a distinction between living Kant’s ideal of a Kingdom of Ends and striving for that ideal, we may in fact intervene in the suicidal plans of others in ways that might initially seem at odds with Kant’s emphasis upon autonomy and choice. While I do not attempt to provide a comprehensive classification of interventions into acceptable and unacceptable, I will try to draw some general conclusions about what a Kantian view of suicide intervention should look like. I should mention at the outset a key caveat. My chief interest will be in suicide attempts that arise from mental illness [4]. While my conclusions bear on suicide unrelated to mental illness, the connection between mental illness and suicide is so intimate that we overlook it to our peril. So unless stated otherwise, references to ‘suicidal persons,’ ‘suicide’ and so on are to cases in which mental illness is a factor. Furthermore, while I may sometimes appear to exaggerate the symptoms and effects of mental illness, this exaggeration has an argumentative purpose: mental illnesses vary in their severity from person to person, but by focusing on the most severe cases and how suicidal behaviour arises in them, I can dramatize the particular effects mental illness has on agents’ psychology and agency.
2. Questions of psychology and rationality loom especially large if we seek to address the moral issues surrounding intervention from a Kantian perspective. For among the prominent contemporary moral theories, Kant’s emphasises above all the centrality of reason. Not only are moral requirements themselves rational requirements in Kant’s moral philosophy, but morality depends for its existence on rational human wills. Rational agency is not only the source of value within Kantian moral theory, but also the essential condition for moral appraisal. For Kantians, moral appraisal focuses on maxims, agents’ justificatory principles of action. Hence, any attempt to understand the psychology underlying suicidal behaviour through a Kantian lens must focus on suicidal agents’ maxims, which are themselves rationally chosen principles. Therefore, any attempt to draw substantive Kantian conclusions about the morality of suicide depends crucially on psychological facts surrounding suicide. Furthermore, Kantian moral theory provides us reasons to intervene when others are intending suicide. First, Kant himself notoriously claimed that suicide is in nearly every instance a violation of a central moral duty, that of self-preservation. While Kant’s arguments for this position are controversial, they are less implausible than they appear [5]. Regardless of this, if suicide is at least sometimes a moral wrong, then this gives us a prima facie reason to act so as to prevent it. In addition, Kant would certainly agree that suicide, even when morally permissible, often represents a profound human and moral tragedy [6]. A life that ends in suicide is a life that ends in sorrow, even if the © Society for Applied Philosophy, 2002
JOA19.3C05
246
11/7/02, 15:47
Suicide Intervention and Non-Ideal Kantian Theory
247
circumstances of the suicide are honourable or even praiseworthy. Suicide necessarily terminates a human life, a life to which Kant attributes a fundamental dignity. Hence, the loss of something of great value seems to justify intervening in suicidal behaviour in many instances. It is far from my purpose to establish when suicide intervention is justified, a question which naturally leads to asking when suicide itself is justified. My concern is rather how we may intervene assuming that we have reasons to do so. But Kantianism at least establishes that we will often have reasons to do so. So presuming we have reasons, moral or otherwise, to intervene in others’ suicidal behaviour on some occasions, and that our interventions may depend for their effectiveness on agents’ psychology, we must consider which psychological facts bear on our attempts at intervention.
3. Kant conceives of the human will as acting on reasons, reasons that agents give themselves in policies of action called maxims. Even an individual’s suicidal intentions can be associated with a maxim, a kind of subjective principle of action stating the agent’s end for acting and her rationale. One such suicidal maxim that especially vexed Kant was the aforementioned “From self-love I make it my principle to shorten my life if its continuance threatens more evil than it promises pleasure [7].” This maxim is important because Kant is particularly concerned to refute a roughly Humean argument that suicide is not unnatural if we are naturally to pursue happiness and to avoid misery. Regrettably, Kant’s preoccupation with this Humean argument leads him to a false picture of phenomenology of suicidal agents. He gives the false impression that many suicides occur when an agent weighs the good of her future happiness against the moral good of respecting one’s own autonomy. Kant depicts the suicidal agent as calmly opting for one good, happiness, over another, rational autonomy. And surely there is a kernel of truth here: by choosing to die, the suicidal agent has ‘renounced’ her rational autonomy. She is willing the will’s demise. But is she renouncing her autonomy in favour of some other “good,” in particular her happiness? A cursory examination of psychiatric and sociological evidence will reveal that Kant’s view oversimplifies the suicidal agent’s psychology: Kant’s picture is that the suicidal agent has misprioritised two classes of goods, placing her own happiness (or at least her release from misery) above the unconditional value of her rational autonomy. But this picture assumes that suicidal agents care about their own happiness and their rational autonomy in the first place. As I shall show, suicidal agents care little for their own happiness, because their state is such that they have come to have a diminished conception of the personal good that constitutes their happiness. Instead of (wrongly) ‘weighing’ two goods, suicidal agents become so disenchanted with the prospect of achieving happiness that they not only give up pursuing happiness, but they repudiate their own conception of what their happiness is. Hence, suicidal agents do not act from a sense of unhappiness but from a kind of nihilism regarding their own personal good. For the benumbed suicidal agent, the world is so drained of value that suicide does not represent the sacrifice of one’s rational autonomy for the sake of one’s own happiness but the disavowal of happiness altogether. © Society for Applied Philosophy, 2002
JOA19.3C05
247
11/7/02, 15:47
248
Michael J. Cholbi
One important parcel of evidence for these claims is that suicidal agents lack the characteristic emotional reactions of individuals concerned with their own well-being. As suicidal persons become more committed to the idea of their own deaths, their capacity for caring about their future diminishes: suicidal agents rarely report the reactive attitudes (regret, frustration, pride, etc.) of those committed to their own happiness [8]. Many suicidal persons even report an absence of, or an indifference to, pleasure and pain. This suggests that at the depth of suicidal depression, we do not experience severe pessimism regarding our hopes for happiness but a kind of apathetic nihilism Pleasure, pain, and these reactive attitudes are the empirical marks of caring about things, [9] but the truly determined suicidal agent is not miserable but withdrawn, unreachable, and seemingly bored with her own existence [10]. Moreover, in comparison with non-suicidal persons, suicidal persons appear to fear death more because of the loss of their social identity and because of the consequences their death would have for their families. This finding, along with behaviours such as arranging one’s affairs and giving away belongings, suggests that the suicidal agent no longer experiences conflicts between her own good and others’ good, since she no longer has a conception of her own good. As Maria von Herbert, a young woman who sought Kant’s counsel regarding her own misery and who took her own life thereafter, expressed it: My vision is clear now. I feel that a vast emptiness extends inside me, and all around me — so that I almost find myself to be superfluous, unnecessary. Nothing attracts me. I’m tormented by a boredom that makes life intolerable. Don’t think me arrogant, but the demands of morality are too easy for me. . . . You can see, perhaps, why I only want one thing, namely to shorten this pointless life, a life which I am convinced will get neither better nor worse [11]. The suicidal agent acts not from self-love but from an abhorrence — note von Herbert’s prediction of “neither better nor worse” — of a disenchanted existence. And while suicide is nearly always associated with depression or a similar mental condition, hopelessness above all characterizes the suicidal individual’s psychology [12]. Eventually, this hopelessness becomes a permanent and constitutional feature of suicidal agents. It manifests itself in a psychological state Edwin Shneidman has termed “psychache.” [13] Psychache is a gripping sensation of “hurt, anguish, and soreness,” a condition rarely found outside depression, schizophrenia or other clinical mental illness. This hopelessness also appears to function as a factor that compromises the ‘psychological’ immune system and makes agents vulnerable to stress. Very often, stressful life events induce panic in mentally ill individuals and precipitate suicide attempts. Not only is suicide often precipitated by setbacks such as the termination of romantic relationships, career failures, confrontations with the law, addictive relapses, etc., but also by positive events such as being released from an institution or initiating a promising course of drug treatment [14]. Suicide may therefore be seen as a symptom of an assault on an already weakened psyche. Much as patients with immune disorders such as AIDS are especially vulnerable to infection, so too are mentally ill patients susceptible to stresses due to their compromised psychological “immune system”; suicide attempts are a common symptom of such immunological assaults. Due to entrenched hopelessness and disengagement, the mentally ill mind lacks the flexibility © Society for Applied Philosophy, 2002
JOA19.3C05
248
11/7/02, 15:47
Suicide Intervention and Non-Ideal Kantian Theory
249
and adaptability to withstand the stresses of life events. Stressful events, by providing apparent grounds for caring about one’s own well-being, confound the agent’s hopelessness. As a result, the act of suicide is often impulsive and poorly thought out, reflecting the intense psychological vulnerability of suicidal persons and their proclivity toward volatility and agitation [15]. Hence, to describe suicide as motivated by unhappiness may be misleading. To be sure, suicidal individuals with depression or other mental illnesses suffer a distinct kind of misery, one that I have argued is closer to indifference to their own well-being rather than pessimism about its realization. But suicidal agents are not acting from a sense of unhappiness, at least as that might be understood in Kantian terms. They do not seek to satisfy the set of inclinations that constitutes their happiness, and hence they are not moved by Kantian “self-love.” Rather, their suicidal conduct originates in the realization that they will probably not recover a conception of their own happiness and the practical agency that accompanies it. The kind of misery that sets in with this realization is undoubtedly profound, but it is distinct from mere pessimism about the prospects for happiness, even if it would be described as such by other non-Kantian views that construe happiness more in terms of psychological sensations. The basis of suicidal hopelessness, I contend, is that the agent’s capacity to care about her own happiness has been nearly ground into submission. To put the point in Kantian language, maxims are not just descriptions of what one is doing. Maxims express the agent’s principle in acting as she does. In the case of many suicidal agents, it may be correct to say, with Kant, that they seek to end their misery, but this neglects the fact that their proper end is not to end their misery but to end their sense of normative listlessness. It as if these agents’ underlying maxim has the end of giving up on having happiness altogether, since no conceivable conception of happiness seems worth pursuing. Kantians may find the concept of such a maxim unintelligible. Can an agent really act on a maxim whose end is that she relinquish her rational power to set ends? That may sound incoherent to some Kantian ears, since all maxims have a determinate end, but the end of this maxim is to give up on end-having by killing oneself. Admittedly, describing the suicidal agent’s maxim is challenging. The oft-heard metaphor of the person as simply a ‘shell’ surrounding a deadened will suggests the maxim might run as follows: Because I no longer can value ends, I take my own life in order to be rid of my body, the instrument of desire and of the pursuit of ends — so as to bring my physical being in line with my psychological being.
4. Scientific research thus suggests that, for the most part, Kant’s concern about agents being moved to commit suicide in order to increase their future happiness is misplaced. Agents are not balancing ends, whether moral or non-moral, when they contemplate suicide. Indeed, their very ability to contemplate suicide appears to arise from a diminished capacity to value ends of their own. Indeed, at its nadir, the suicidal agent suffers a loss of self or of practical identity. A commitment to the personal ends that comprise one’s happiness is an essential element of autonomous rational agency, even for a philosopher like Kant who relegated © Society for Applied Philosophy, 2002
JOA19.3C05
249
11/7/02, 15:47
250
Michael J. Cholbi
happiness to a subsidiary role in his theory of value. Kant was wrong to describe the suicidal agent as a wanton, ‘giving in’ to her own happiness or desires. For in order to give in, the agent must possess a sense of normative identity, the rejection of which constitutes giving in. But a suicidal agent very often does not have the sorts of evaluative commitments to ends and personal projects that serve to orient her deliberation and ground her identity as a practical agent in the first place. This relative absence of evaluative commitments, incentives and desires — the constituents of “the will to live” — threatens or erodes the suicidal agent’s rational agency. When we deliberate in the way Kant supposes, we do not simply act upon some desire or other. We deliberate as if there is some unified being whose voice we call the will, a self over and above the considerations that move us to act and that serves as the author of our maxims and actions. This self is largely defined by a set of evaluative commitments the pursuit of which is the pursuit of happiness. But our acts and reasons can only be reflections of the Kantian will if there is a distinct self, the standpoint of the will, that engages in deliberation. The suicidal person is an individual without a practical identity and hence without an important background condition for Kantian rationality. This results from a slow erosion of her practical identity and eventually, the realization that her own rational agency is no longer worth anything to her, for nothing has value to her [16]. The absence of a practical identity and its accompanying incentives and evaluative commitments is revealed in suicidal individuals’ feelings of a self at odds with itself and a world that is recalcitrant to their pursuits. Suicidal individuals in fact report that they feel just frustrations of agency. For instance, empirical research vindicates M.L. Farber’s claim [17] that suicidal hopelessness tracks agents’ sense of self-competence, their sense of being able to control the environment in ways adequate to meet their wishes and needs. This confirms the proposed Kantian account of the psychology of suicide in that it reflects an extreme disenchantment with the world and with the agent’s pursuit of her happiness. Another psychological phenomenon that reflects the absence of a substantive practical selfhood in suicidal persons is their frequent self-object confusions. That is, the suicidal agent often harbours feelings of self-hatred or hopelessness that are then projected onto others and onto the world [18]. This self-object blurring has a number of manifestations. Some suicidal agents deny responsibility for their suicidal intentions, claiming that Fate or Nature determines their actions, [19] or they blame their therapists [20]. This apparent destruction of one’s practical self also finds vivid expression in suicidal persons’ occasional sense of disembodiedness or fission. Many suicidal persons experience their minds and bodies as disconnected, as if their consciousness is the terminus of two distinct causal series, one mental and one bodily. Such individuals may identify one of these as the “killer” and the other one as ‘the self,’ though the killer and the self are both understood as “me.” [21]. Randall Jarrell [22] said of his own suicidal manic depression that it was as if his normal self was stolen away and replaced with a log, as if he were only a husk of his former self. Thus, as Barbara Herman has pointed out, when we no longer seek satisfaction from our existence — when we, like suicidal persons, no longer pursue happiness nor avoid misery since we no longer believe the world is suited to satisfying our desires and inclinations — we quickly lose our self-conception as unified agents. As both rational and imaginative creatures, © Society for Applied Philosophy, 2002
JOA19.3C05
250
11/7/02, 15:47
Suicide Intervention and Non-Ideal Kantian Theory
251
we not only can envisage what it would be like to do one thing or another, we recognize that we have a future whose shape we can affect. We come to have a concept of a life — something that will connect our experiences — and about which we feel satisfaction, or not. [23] Suicidal persons are so overcome by their misery that no personal ends are worth pursuing and so they no longer have a concept of a life, their life, that connects their experience and which can go well or ill [24]. They are no longer complete agents in a Kantian sense, for they lack a key ingredient of Kantian agency and a central drive toward self-individuation. Thus, in the sort of suicide that especially concerned Kant, the usual psychological story is much more complex than he supposed. Agents do not simply calculate future goods and evils and opt for the good of death over the evils of life. Instead, the evils of life accumulate until the point that agents no longer feel a sense of attachment to their own happiness at all and are indifferent to their existence. It is here that a necessary condition of willful Kantian agency and of rational deliberation — a backdrop of evaluative commitments, incentives, and projects — is most threatened [25].
5. None of this is to say that suicidal agents are aware of or can express these facts about their diminished rational agency with any degree of philosophical precision. But I wish to argue that this revised neo-Kantian understanding of the psychology of suicide has important implications for the morality of suicide intervention. There are a number of forms that suicide intervention may take. We might point out to the suicidal person facts that warrant optimism about her future, facts she had previously neglected. We might try to direct the suicidal agent to medical or psychological care for her condition. We could point out the number of people who depend on her and who would suffer from her death. We could actively seek to alleviate the sources of her discontent, by paying off her debts, finding her a more satisfying job, introducing her to our friends, and so on. Insofar as these do not appear to violate the autonomy of the agent or fail to treat her as an end in itself, they appear to be plausible candidates for acceptable forms of suicide intervention. But note that even if such actions would be permissible on Kant’s view, our revised Kantian psychology of suicide illustrates how ineffective such measures are likely to be. In each of these forms of intervention, its effectiveness depends upon the agent’s being susceptible to considerations or changes in her life that further her ends or make her happier. We seek to make her care about ends like community approval, camaraderie, vocational success, and the like. But if Kant is incorrect and suicide is a desperate act rather than a calculated one, these approaches are almost certainly doomed to fail, for they depend on the presence in the agent of standing commitments or ends that genuinely suicidal persons lack. It has been recognized since Durkheim that most suicidal persons are socially isolated, unable to form lasting relationships with others and often unable to express their emotional needs and dependencies to others, despite the likelihood that their emotional needs are more intense than others’. Suicidal persons’ social isolation only sharpens the point that the more benign, non-coercive methods of intervention © Society for Applied Philosophy, 2002
JOA19.3C05
251
11/7/02, 15:47
252
Michael J. Cholbi
are unlikely to reach them, for agents without a rich network of social ties are unlikely to be persuaded by such methods. Moving agents by appealing to ends such as happiness or duty to others when they lack ends is therefore exceedingly difficult. Of course, the fact that these forms of intervention are probably ineffective does not demonstrate that we are not obliged to pursue them. Some obligations call for heroic but hopeless behaviour. But this fact implies that a Kantian view of suicide intervention should at least advocate early intervention, before the agent’s autonomy is so curtailed that her responsiveness to reasons based in particular ends is eliminated. We must reach the would-be suicidal person before her capacity to respond to particular ends is depleted by her sense of the world as an evaluative vacuum.
6. When early intervention is not possible or is simply not undertaken, we are confronted with an instrumental dilemma, since seriously suicidal persons often make every effort to make intervention impossible. But a moral dilemma arises too. Struck by the prospect of someone close to us pondering suicide, many of us will undoubtedly contemplate more drastic measures. When an agent has fully committed herself to suicide, the above methods, which rely on persuasion and the agent’s attachment to ends, are almost certainly going to be ineffective. We might instead lie to a seriously suicidal person, not only about our opinions but about matters of fact; we might offer promises that we may not intend nor be able to keep; we might place psychotropic drugs in the suicidal person’s food; we might place her in an institution against her will. All of these methods are likely to be more effective than appeals to reason, as the likelihood of the suicidal act grows. But from a Kantian standpoint, such measures appear morally dubious. They involve deceit, manipulation, denial of consent, and thwarting of others’ expectations, all of which fail to treat the agent fully as an end in herself, and Kant is notoriously hardheaded and unyielding about conduct, like lying, that disregards others’ rational autonomy. Of course the non-persuasive forms of intervention are motivated by the desire that the agent be returned to a position of dignity — that she be able to treat herself as an end. Though we are not treating her as a means only, not every act meant to treat others as ends in themselves in fact treats them as such. Were we to engage in these non-persuasive forms of suicide intervention, our maxims would seem to be internally at odds [26]. Such tradeoffs might be easy to accommodate within a consequentialist moral framework, but if the non-persuasive forms of suicide intervention are not allowed on a Kantian view, this is because they amount to treating the suicidal agent as something close to a child [27]. And in the cruellest of ironies, we repudiate the agent’s autonomy in order to save it. We are therefore in something of a practical dilemma: moral permissibility and instrumental effectiveness are inversely related. So must Kantian-minded moral agents refrain from intervention in serious cases? I believe the answer is ‘no’. As a person’s practical identity deteriorates, their status as a citizen of Kant’s Kingdom of Ends is threatened, and we are licensed to undertake otherwise unjust or immoral measures to halt this deterioration with the hope that they may reacquire a practical identity, a set of normative commitments, and their place within the Kingdom of Ends. It certainly cannot be the case that we have less reason to © Society for Applied Philosophy, 2002
JOA19.3C05
252
11/7/02, 15:47
Suicide Intervention and Non-Ideal Kantian Theory
253
intervene as suicidal intent grows. But in order to make the case for the legitimacy of non-persuasive interventions, we must first recognize the special circumstances that a possible suicide presents to us. Kant’s ethics presents us with an ideal of human relations, the Kingdom of Ends, in which each individual is valued as an end in herself and is treated as such. Critics sometimes charge that Kant makes overly idealistic assumptions that oversimplify our moral obligations when the Kingdom of Ends is not at hand. One such assumption is that the agents affected by our behaviour are psychologically normal and hence are full-blown agents with rational autonomy. As we have seen, this is not true of the great majority of suicidal agents, and even when it is false and suicidal behaviour is unrelated to mental illness, it can be difficult, especially for non-professionals, to determine if a suicidal person is mentally ill. Hence, not only must we often act under uncertainty, but we are contemplating intervention in response to a non-ideal circumstance, namely, the diminished rational autonomy of the suicidal individual. We can therefore ask what Kant’s ideal requires of us in non-ideal circumstances and hence see our way to a resolution of the dilemma mentioned above — that in the cases where we most have reason to intervene in others’ suicidal plans, the means by which we may intervene are likely to be either instrumentally ineffective or morally impermissible.
7. The notion that we may need a special understanding of Kantian morality for nonideal circumstances is due to John Rawls [28]. In nonideal circumstances, we do not spurn the values we adhere to in ideal circumstances; rather, the ideal theory and its value remain fundamental in that the “ideal conception shows how the nonideal scheme is to be set up.” [29] While the main nonideal circumstance that concerns Rawls is agents not complying with the imperatives of justice, he acknowledges that the “unjust designs” of certain agents are not the only nonideal circumstance [30]. Natural limitations and historical contingencies also may demand the use of nonideal theory. We assume in ideal theory that persons are rational and able to manage their own affairs, but nonideal circumstances may demand “adjustments to the natural features of the human situation, as with the lesser liberty of children.” [31] In such situations, we may act in paternalistic ways that would be disallowed in ideal circumstances, on Rawls’ view. The need for paternalistic principles arises from exactly those situations where agents may lack their rational capacities. We have reason to accept these principles, Rawls argues, because they are a “protection against our irrationality.” [32] I contend that the best response Kantians can offer to the dilemma I have identified must begin with the recognition that the suicidal plans of others, especially of the mentally ill, are an instance of nonideal conditions. In particular, the dilemma weakens if we sometimes treat Kant’s Kingdom of Ends not as an ideal to be lived directly but as a goal to be realized as best we can in nonideal conditions. That is, we are “to work toward the conditions in which [the Kingdom] is feasible,” [33] which may sometimes require that we spurn the Kingdom as an ideal. To act so as to honour Kant’s ideal may after all undermine it when natural or moral evil intercedes. “To act as a member of the Kingdom can make the Kingdom more, and not less, remote.” [34] When acting as a member of the Kingdom would make it more remote, Kantianism can be © Society for Applied Philosophy, 2002
JOA19.3C05
253
11/7/02, 15:47
254
Michael J. Cholbi
understood to demand of us that we act as if we occupied a place in the Kingdom of Ends. Given the moral and natural imperfection of our world, much of our conduct is by necessity ‘as if ’ conduct, for acting as if the ideal of the Kingdom of Ends is in place when it is not is frequently irrational and immoral. So what our conduct should strive for is the endorsement of Kant’s ideal rather than its direct embodiment. When we lie to keep the Nazis at bay in order to save lives, or we break promises in order to prevent unjust treatment of others, we honour Kant’s ideal, not by living it (for the presence of evil circumstances precludes that) but by striving toward the ideal [35]. But by utilizing the difference between living the ideal and striving for it, we can sidestep the conclusion of the previous section, namely, that many measures likely to be effective in stopping the mentally ill from committing suicide are morally impermissible. Going on strict considerations of duty, the intervener cannot trample on the agent’s autonomy. Yet if the intervener really wishes to treat the suicidal agent as an end in itself — if she really cares about her, her humanity, and her dignity, if she loves her as a person — then the intervener must act in ways that encourage this dignity and encourage the possibility of treating this agent as an end in the future. The agent may permissibly ignore otherwise binding Kantian duties, not because she rejects Kant’s ideal of the Kingdom of Ends, but precisely because she accepts it wholeheartedly even in the face of a natural evil such as the self-destruction of another rational will. Kant may have been wrong about suicide debasing our humanity, but as I have shown, mental illness actually debases the humanity and agency of its victims.
8. But does not this move give the agent too much leeway, allowing her to perform any form of intervention she thinks effective even when such intervention is wildly at odds with the autonomy or dignity of the suicidal agent? Should we be ready to institutionalize our loved ones at the slightest signs of depression? My resolution, it may be argued, is far too utilitarian to provide a Kantian answer to our dilemma [36]. As with much in Kant’s moral philosophy, I doubt that a comprehensive and watertight casuistry can be provided on this question. The most we can recommend is some sense of proportionality between the seriousness of the suicidal individual’s intentions and the means needed to discourage her. With persons suffering from a transient case of the blues, the persuasive methods, methods exemplified in psychotherapy, are obviously best. But as mental illness deepens, so too should our willingness to deploy non-persuasive methods, methods that do not involve giving the agent good reasons that she then adopts, but methods that may involve subtle forms of reticence, duplicity, coercion, and control. This does not entail that the situation may become severe enough that the most heinous means are called for; killing X to prevent Y’s suicide is unfathomable on any understanding of the Kingdom of Ends. In adopting the Kingdom of Ends as an ideal to be encouraged and endorsed instead of lived and directly realized, we do not thereby adopt the doctrine of “by whatever means necessary.” So in adopting proportionality we do not warp Kantianism into consequentialism. The agent’s own autonomy sets limits on how we may permissibly intervene in her suicidal plans. The best we can do is to envision ourselves as the object of a paternalistic intervention and ask ourselves if we would have reason to accept that intervention after the fact and © Society for Applied Philosophy, 2002
JOA19.3C05
254
11/7/02, 15:47
Suicide Intervention and Non-Ideal Kantian Theory
255
whether we might regret others not intervening on our behalf. However we intervene, it must be possible for us to envision the suicidal agent as later offering her rational consent to the intervention. We must, as Rawls put it, “be able to argue that with the development or recovery of his rational powers the individual in question will accept our decision on his behalf and agree with us that we did the best thing for him.” [37] Such a criterion reflects the thought that the agent’s autonomy is to be respected by proxy. Should we have reason to think that the agent would not agree to have her liberty restricted in a particularly coercive way, then this reason restricts how we may restrict that liberty. Some mentally ill individuals have even made possible such protection by proxy with so-called “Odysseus arrangements,” advanced instruction directives stipulating beforehand which treatments they consent to receive in response to future suicidal episodes [38]. Such arrangements aid us in restoring autonomy in the least unjust way. Many may find my recommendations of early intervention plus proportionality too obvious. I grant that these are fairly benign recommendations. Yet one of moral philosophy’s roles is to provide the detailed arguments and rich conceptual background needed to bolster our considered moral judgments. Moreover, given the often shameful history of mental health care, in which proportionality in particular was disregarded, my recommendations are clearly not obvious to some. For instance, my recommendations would render a split decision on the experience Susanna Kaysen describes in her memoir Girl, Interrupted. Two years after an apparent suicide attempt in which she consumed large quantities of aspirin and vodka, Kaysen was sent to a psychiatrist. After a single session, the psychiatrist drew upon the suicide attempt and Kaysen’s general moodiness and lack of motivation to diagnose her with the notoriously slippery “borderline personality disorder”. She was sent to a psychiatric hospital, where she remained for two years. Kaysen’s family and the psychiatrist, despite their questionable motives, are to be lauded for their early intervention but condemned for the disproportionality and heartlessness of their response. At the same time, our culture’s sense of shame at mental illness and our ignorance about it too often delay intervention [39]. As Kaysen herself notes, the transition from ‘normal’ to mentally ill is often gradual: “Most people pass over incrementally, making a series of perforations between here and there until an opening exists.” [40] How many times have we heard tales of suicides in which everyone misinterpreted the clues or unconsciously repressed a growing crisis? ‘No one saw it coming’ is too common a refrain.
9. Kant’s moral philosophy, despite its emphasis on freedom, is therefore not completely hostile to paternalism with respect to suicide intervention [41]. If I am correct in my proposal for a Kantian understanding of suicide intervention, then we may sometimes act paternalistically in the interests of the rational agent in order to salvage the prospect of her rational autonomy [42]. Such a position is nonetheless recognizably Kantian in three respects: (a) The agent’s rational autonomy is the guiding value to consider when addressing intervention. (b) The insistence on proportionality imposes limits to how we may trade off or exchange existing autonomy for long-term autonomous agency. (c) Individuals’ lives are valued, and intervention justified, in terms of rational autonomy, not in terms of the potential happiness of the agent. © Society for Applied Philosophy, 2002
JOA19.3C05
255
11/7/02, 15:47
256
Michael J. Cholbi
Of course, I have not fully resolved the dilemma I outlined in section 4. There is probably no Kantian approach to suicide intervention that will demonstrably license all and only those forms of suicide intervention that are both effective and uncontroversially morally permissible. But at least early intervention coupled with a sense of proportionality between the means of intervention and the seriousness of the suicidal intent is a start. Moreover, my conclusions can be applied not only to suicidal behaviour associated with mental illness but to other cases as well. First, despite the strong empirical association between mental illness and suicide, we may be uncertain as to whether a suicidal individual is fully autonomous or is suffering from a mental illness that has diminished her rational agency. However, this uncertainty seems to justify a rebuttable presumption that any instance of suicidal behaviour has its origins in mental illness. Assuming that mental illness underlies suicidal behaviour and thereby intervening may violate the autonomy of agents not suffering from mental illness, but assuming the contrary would preclude intervention even when the agent is not acting with full rational autonomy. Agents autonomously choosing to die can of course do so after our attempts at intervention, but agents suffering from a lack of rational autonomy do not have the opportunity to reconsider their suicides after the fact. This asymmetry makes it reasonable to operate on the presumption that evidence of suicidal ideation is also evidence of mental illness. In addition, proportionality is a sound principle even when we have little reason to think that the suicidal agent is suffering from a mental illness and its attendant loss of practical agency. For agents who are in fact fully rational and autonomous but about whom we are uncertain as to their mental condition, proportionality ensures that we do the least harm or violate their autonomy in the least troubling ways. At the same time, how to treat the fully autonomous suicide indicates where a Kantian position reaches its self-imposed limits. When suicide is fully autonomous, intervention appears morally unjustified, and this is so even if we agree that such suicide represents an undesirable way for life to end. Whether intervention is justified on grounds unrelated to the agent’s autonomy (the feelings of the agent’s loved ones, for example) is not a question I shall pursue. Likewise, where there is no chance of ever restoring an individual’s practical agency, intervention inflicts a kind of cruelty, forcing the agent to suffer. Both cases illustrate how a Kantian morality must tolerate tragedy [43]. To sum up, I have to sought to address whether we may, from a Kantian point of view, intervene in the suicidal plans of mentally ill persons, and if so, how. Though we can understand the psychology of suicide as resting on the absence of conditions for unified and autonomous rational agency, this in turn makes many effective means of suicide intervention morally impermissible. Thus, it would seem that in the cases where we most have reason to intervene, we either may not do so or may only do so ineffectually. However, since the suicidal plans of the mentally ill represent nonideal circumstances, we may act in ways that further their autonomy even if these actions are coercive, duplicitous, or manipulative. Michael J. Cholbi, Department of Philosophy, Brooklyn College, City University of New York, 2900 Bedford Ave., Brooklyn NY 11210, USA. Fax: (718) 915-4675.
[email protected] © Society for Applied Philosophy, 2002
JOA19.3C05
256
11/7/02, 15:47
Suicide Intervention and Non-Ideal Kantian Theory
257
NOTES [1] L. F. MC and B. J. S, National hospital ambulatory care survey, 1996. Emergency department summary advance data from vital and health statistics, no. 293. (Hyattsville, Md., National Center for Health Statistics, 1997). [2] St. T A, Summa Theologica II.Q64.A5, and D H, Of suicide, in Miller (ed.), David Hume: Essays Moral, Political, and Literary, rev. ed. (Indianapolis, Liberty, 1985), 577–589, are typical examples of philosophers interested entirely in this question, as is Kant. [3] In one landmark study, 98% of those who died by suicide were found to be clinically mentally ill, with 94% suffering from serious psychiatric illness, and 68% suffering from manic depression. See E R, G E. M, et al., Some clinical considerations in the prevention of suicide based on a study of one hundred thirty-four successful suicides, American Journal of Public Health 49 (1959), 888– 89. The close link between suicide and depression and other mental illnesses has been confirmed by T. L. D and H. S. R, A study of suicide in the Seattle area, Comprehensive Psychiatry 1 (1960), 349–59; by B. B, B. J. B, et al., A hundred cases of suicide: clinical aspects, British Journal of Psychiatry 125 (1974), 355–73; by D. W. B, The Iowa record-linkage experience, Suicide and Life-threatening Behavior 19 (1989), 78–89; and by J. A, F. A, and H. H. S, Suicide risk in patients with major depressive disorder, Journal of Clinical Psychiatry (Suppl. 2) (1999), 57–62. V S (1996), Mental Illness, (Lucent) has concluded that about 15 percent of people who seek treatment for depression commit suicide. Admittedly, such studies must be carefully constructed so that suicide attempts as not construed as evidence for mental illnesses, thus biasing the evidence in favour of a connection between suicide and mental illness. [4] For a taxonomy of the varieties of suicide, see G J. F (1995) Contemplating Suicide: The Language and Ethics of Self-harm (New York, Routledge), ch. 9. [5] See my Kant and the irrationality of suicide, History of Philosophy Quarterly 17:2 (April 2000) for a thorough discussion of Kant’s arguments. [6] See T E. H, Jr. (1991) Self-regarding suicide: a modified Kantian view, in his Autonomy and Self-Respect (Cambridge), pp. 85–103. [7] Groundwork of the Metaphysics of Morals, 422. Citations are to the Berlin Akademie edition pagination and Mary Gregor’s translations in (1996) Kant’s Practical Philosophy (Cambridge, Cambridge University Press). [8] The remarks in this and the subsequent paragraph summarise R A. B (1999) Essentials of Psychology, 2d ed. (Allyn & Bacon, Needham Heights, Mass.), 471–72. [9] B H, Transforming moral incentives, unpublished manuscript. [10] This sense of ennui, emptiness, and fatigue comes through in many suicide notes. See A. A. L (1988) Suicide Notes: Predictive Clues and Patterns (New York: Humans Sciences). [11] From H to Kant, January 1793, in Zweig (ed.) (1967) Kant: Philosophical Correspondence, 1759–1799 (Chicago, University of Chicago Press). [12] A T. B, M K, and A W, Hopelessness and suicidal behavior, Journal of the American Medical Association 235 (1975), 1146–49, and J. F, W. A. S, et al., Time-related predictors of suicide in major affective disorder. American Journal of Psychiatry 147 (1990), 1189–1194. [13] E S, Suicide as psychache, Journal of Nervous and Mental Disease 818 (1993), 147–49. [14] See E. S, A. D, and C. C, Suicide upon recovery from depression: a clinical note, Journal of Nervous and Mental Disease 176 (1988), 633–636; H. G. M and P. P, Suicide and other unexpected deaths among psychiatric inpatients, British Journal of Psychiatry 158 (1991), 368–74; and K. A. B, D. C. C, J. F, and H. M. K, Clinical features of inpatient suicides, Psychiatric Annals 23 (1993), 256–62. [15] C. L. W, J. A. D, and I. M, Impulsive suicidal behavior, Journal of Clinical Psychology 36 (1980), 90–94; A. A, R. P, and H. M. P, Anxiety, impulsivity and depressed mood in relation to suicidal and violent behavior, Acta Psychiatrica Scandinavica 87 (1993), 1–5; and P. N, P. G, G. E, and M. A, Temperamental vulnerability and suicide risk after attempted suicide, Suicide and Life-threatening Behavior 26 (1996), 380–94. [16] B H, in Transforming moral incentives, argues forcefully for the view that pursuing our happiness, and thus valuing ourselves as beings with desires, inclinations, and evaluative commitments, is a necessary condition for both full rational agency and self-individuation. See also R D © Society for Applied Philosophy, 2002
JOA19.3C05
257
11/7/02, 15:47
258
[17] [18]
[19] [20] [21] [22] [23] [24]
[25]
[26]
[27]
[28] [29]
[30] [31] [32] [33]
Michael J. Cholbi (1993) Life’s Dominion (New York, Knopf ), and A J (1999) Respecting the margins of agency: Alzheimer’s patients and the capacity to value, Philosophy and Public Affairs 28, 105–38, for similar views about valuing and agency. M. L. F (1967) Theory of Suicide (New York, Funk & Wagnalls). C. V. L hypothesizes that this self-object blurring may lead to suicidal thoughts under three conditions: when persons fail to forge a separate identity from their parents, when persons rely nearly exclusively on external controls (social disapproval, e.g.) to regulate their behaviour, or when persons lack a diversity of problem-solving techniques and rely on inflexible problem-solving approaches. See his (1967) Understanding and Preventing Suicide (Springfield, Ill., Charles C. Thomas). S S. A, Suicide, and the hidden executioner, International Review of Psychoanalysis 7 (1980), 51–60. H H, Psychotherapy and suicide, American Journal of Psychotherapy 35 (1981), 469–80. J T. M and D H. B, The devices of suicide: Revenge, riddance, and rebirth, International Review of Psychoanalysis 7 (1980), 63. M. J (ed.) (1985) Randall Jarrell’s Letters: An Autobiographical and Literary Selection, (Boston, Houghton Mifflin). Herman, Transforming incentives, §3. My claim is not that the suicidal individual’s practical identity is threatened because she is unable to consider the narrative of her life from an evaluative point of view. Rather, the individual is unable to see the present or future as imbued with value, even if she attaches normative significance to her past. See Jaworska (op. cit. p. 119). I therefore break with E M regarding the way mental illness distorts the reasons on which suicidal choices are often made. See his Choosing death: philosophical observations on suicide and euthanasia, Philosophy, Psychiatry, and Psychology 5 (1998), 108–09. Matthews seems to have in mind that mental illness makes certain reasons more compelling than they really are, whereas the rationality I have in mind is systemic, having to do with the conditions for choice instead of the character of the agent’s apparent reasons. Fairbairn (1995, pp. 72–73) makes a similar error regarding rationality and autonomy. A surprising consequence of this disenchantment from the possibility of happiness is that suicidal agents may find morality very unchallenging. Maria von Herbert’s remarks indicate that her will is pure, untainted by inclinations or incentives that may lead her morally astray. Somewhat paradoxically then, suicidal agents may be better situated to engage in morally exemplary conduct, despite their diminished autonomy. In his more moralistic moments, Kant claims that we should be happy to be rid of such inclinations or incentives. Yet he also acknowledges not only that we can and ought to direct these incentives and inclinations in the direction of virtue, but that incentives, inclinations, and standing value commitments are needed for imperfectly rational beings such as ourselves to kickstart moral deliberation. Note that the incoherence of this maxim mirrors the incoherence Kant attributes to the morally impermissible maxims of suicidal agents: just as the rational agent pondering suicide is autonomously considering a maxim that fails to give due accord to the value of her rational autonomy, so too do we fail to treat the autonomous agent as fully autonomous in the hope that she will come to treat herself as autonomous and worthy of care. Suicidal agents may, like children, be “passive citizens” of the Kantian Kingdom of Ends who, while faced with the problem of how to work out a plan for conducting themselves, lack an established constitution to guide their wills in a lawlike manner. See Tamar Schapiro, “What is a child?” Ethics 109 (1999): 715–38. Herman, “Transforming incentives,” suggests a similar point with her claim that children lack a full imaginative capacity to formulate a conception of the good. Obviously, this comparison is somewhat misleading. In the case of children, their lack of rational agency is presumably only temporary, whereas that same lack could well be permanent in the mentally ill. See also note 41 below. J R (1971) A Theory of Justice (Cambridge MA, Harvard/Belknap). Hereafter TJ. TJ 241. For some doubts about attempts to base the nonideal principles on the principles of ideal theory, see M P, Reflections on the transition from ideal to nonideal theory, Nous 19 (1985), 551–70. TJ 243. TJ 244. TJ 250. Christine M. Korsgaard, The right to lie: Kant on dealing with evil, in her (1996) Creating the Kingdom of Ends (Cambridge, Cambridge University Press), 148.
© Society for Applied Philosophy, 2002
JOA19.3C05
258
11/7/02, 15:47
Suicide Intervention and Non-Ideal Kantian Theory
259
[34] R L, Duty and desolation, Philosophy 67 (1992), p. 504. [35] Korsgaard, in The right to lie, p. 152, maintains that suicide itself may well be justified in nonideal circumstances of “very great moral or natural evil.” [36] Langton, Duty and desolation, p. 502, doubts that the ideal/nonideal distinction can avoid collapsing nonideal theory into consequentialism. [37] TJ 249. [38] K R J (1999) Night Falls Fast: Understanding Suicide (Knopf ), 261. [39] Jamison vividly depicts the common reticence, shame, and collective paralysis about suicide that even seeps into the mental health community in Jamison (1999) and (1995) An Unquiet Mind: A Memoir of Moods and Madness (Knopf ). [40] S K (1993) Girl, Interrupted (Vintage, New York), 5. [41] I am using “paternalism” here simply to designate apparently unjustified coercion, manipulation, denial of autonomy, etc. One might after all hold that the dilemma I have identified does not arise in some situations that call for acts we would term paternalistic. For instance, one might argue that parents may act in paternalistic ways to prevent their children’s suicide and that this raises no dilemmas about the permissibility of so acting, owing to the special characteristics of the child-parent relationship. I am happy to grant that the dilemma I identify may itself vary in its stringency based on the nature of the relationship between the suicidal individual and the intervenor. [42] Dworkin comes to a similar conclusion about those with conditions such as Alzheimer’s. See Dworkin, Life’s Dominion, ch. 8. [43] See H, Self-regarding suicide: a modified Kantian view, p. 89.
© Society for Applied Philosophy, 2002
JOA19.3C05
259
11/7/02, 15:47
