A Right to Choose Death?
A moral argument for the permissibility of euthanasia and physician-assisted
F. M. Kamm
Do people have a right to choose death? More particularly, are euthanasia
and physician-assisted suicide morally permissible? To clarify terms:
Euthanasia involves a death that is intended to benefit the person who dies,
and requires a final act by some other person (for example, a doctor); physician-assisted
suicide, which requires a final act by the patient, can also be undertaken
for the good of that patient, and I will confine my attention to cases in
which it is. The essential point is that both involve intentionally
ending a human life: In voluntary euthanasia, the patient and doctor both
intend the death; in physician-assisted suicide, the patient intends the death
and the doctor may.1 But how, some ask, can we ever permit people to intentionally
end human lives (even their own lives) without degrading human life? How,
others ask, can we simply prevent people from deciding when to end their own
lives without denying people the autonomy so essential to the value of a human
life? As this pair of questions suggests, the debate about the right to choose
death may appear to present a stand-off between people who endorse life's
intrinsic value, and those who think life's value depends on the interests,
judgments, and choices of the person whose life it is.
This picture of irreconcilable moral conflict is, I believe, too despairing
about the powers of moral argument. To make headway, however, we may need
to pay closer attention to the complexities of cases and the specific moral
terrain they occupy: to think about people on medication, being treated by
physicians, sometimes relying on technical means to stay alive, trying to
decide how to live out what remains of their lives. I will explore this terrain
in moral, not legal, terms: I will be asking you to consult your moral
judgments about cases, and follow out the implications of those judgments.
Though this moral argument bears on constitutional argument and on appropriate
legislation, I will not propose laws or rules for judges, doctors, or hospital
administrators to consult, or worry about slippery slopes created by legally
hard cases. The moral landscape affords firmer footing, and does not, I will
suggest, permit a blanket ban on euthanasia and physician-assisted suicide:
Though both involve intentionally ending human lives, both are sometimes morally
permissible. I will conclude by discussing a different argument for such permissibility
offered by a distinguished group of moral philosophers in a recent amicus
brief to the Supreme Court.
I. Logical Troubles?
Before getting to the issue of moral permissibility, we need to overcome
a preliminary hurdle. I said that euthanasia and physician-assisted suicide
are intended to benefit the patient. Some may object that these ideas make
no sense. How is it possible for death to benefit the person who dies? Death
eliminates the person--how can we produce a benefit if we eliminate the potential
To see how, consider the parallel question about death as a harm: Can a person
be harmed by her own death even though death means that she is no longer around
to suffer the harm? Suppose Schubert's life would have included even greater
musical achievement had he not died so young. Because musical achievement
is an important good, Schubert had a less good life overall than he would
have had if he lived longer. But living a less good life is a harm. By excluding
those achievements, then, Schubert's death harmed him: it prevented the better
life. Now come back to the original concern about how death might be a benefit.
Suppose a person's life would go on containing only misery and pain with no
compensating goods. That person will be better off living a shorter life containing
fewer such uncompensated-for bad things rather than a longer one containing
more of them. But living a better life is a benefit. By interfering with the
important bads, the person's death benefits him: it prevents the worse life.
It is possible, in short, to benefit a person by ending his life. The concept
of euthanasia is, therefore, at least not simply logically confused; similarly
for the idea that physician-assisted suicide may be aimed at the good of the
patient. But conceptual coherence does not imply moral permissibility. So
let's turn now to the moral question: Is it ever morally permissible to benefit
a person by hastening his death, even when he requests it?
II. A Right to Choose
Suppose a doctor is treating a terminally ill patient in severe pain. Suppose,
too, that the pain can only be managed with morphine, but that giving the
morphine is certain to hasten the patient's death. With the patient's consent,
the doctor may nevertheless give the morphine. Why so? Because, in this particular
case, the greater good for the patient is relief of pain, and the lesser evil
is loss of life: after all, the patient is terminally ill, and in severe pain,
so life would end soon anyway and is not of very good quality. So the patient
is overall benefited by having a shorter pain-free life rather than a longer,
even more painful life. (Notice that this could be true even if the morphine
put the patient in a deep unconscious state from which he never awoke, so
that he never consciously experienced pain-free time.)
In giving morphine to produce pain relief, the doctor foresees with certainty
(let's assume) that the patient will die soon. Still, death is a side-effect
of the medication, not the doctor's goal or reason for giving it: the doctor,
that is, is not intending the patient's death, and would give the medication
even if he thought death would not result. (If I have a drink to soothe my
nerves and foresee a hangover, it does not follow that I intend the hangover.)
Because the intended death is not present, we don't yet have a case of euthanasia
or physician-assisted suicide. At the same time, in giving morphine for pain
relief, the doctor is not simply letting the patient die as the disease runs
its course; he administers a drug which causes death. So I think this should
be understood as a case of killing, even though the doctor does not intend
the death. (In other cases we have no trouble seeing that it is possible to
kill without intending death: consider a driver who runs someone over while
Now suppose the morphine loses its power to reduce the intensity of the patient's
pain, but that administering it would still shorten the patient's life and
thus limit the duration of his pain. Suppose, too, that the patient requests
the morphine; fully aware of its effects, he wants to take it so that it will
end his pain by killing him. In short, we now have a case of morphine for
death rather than morphine for pain relief. Is it still morally
permissible to give the morphine? Some people say that we may not kill in
this case. They do not deny that relief of pain is still the greater good
and death the lesser evil: they know that the consequences are essentially
the same as in the case of morphine for pain relief. The problem, they say,
lies in a difference of intent. In the case of giving morphine for pain relief,
we intend the pain relief, and merely foresee the death; but in the case of
giving morphine for death, we intend the death (which is the lesser evil):
we would not give the morphine if we did not expect the death. But some people
think it is impermissible to act with the intent to produce an evil. They
support what is called the Doctrine of Double Effect, according to
which there is a large moral difference between acting with the foresight
that one's conduct will have some evil consequence and acting with the intent
to produce that same evil (even as part of or means to a greater good). So
whereas killing the patient by giving morphine for pain relief is permissible,
killing the patient by giving morphine for death is impermissible.
The distinction between intending an evil and merely foreseeing it sometimes
makes a moral difference. But does it provide a reason to refrain from performing
euthanasia or assisting in suicide? I think not. On many occasions already,
doctors (with a patient's consent) intend the lesser evil to a person
in order to produce his own greater good. For example, a doctor may
intentionally amputate a healthy leg (the lesser evil) in order to get at
and remove a cancerous tumor, thereby saving the patient's life (the greater
good). Or, he may intentionally cause blindness in a patient if seeing would
somehow, for example, destroy the patient's brain, or cause him to die. Furthermore,
he may intentionally cause someone pain, thereby acting contrary to a duty
to relieve suffering, if this helps to save the person's life. The duty to
save life sometimes just outweighs the other duty. Why then is it impermissible
for doctors to intend death when it is the lesser evil, in order to produce
the greater good of no pain; why is it morally wrong to benefit the patient
by giving her a shorter, less painful life rather than having her endure a
longer, more painful one? Recall that in the case of morphine for pain relief,
it was assumed that death would be the lesser evil and pain relief the greater
good. That was one reason we could give the morphine. Why is it wrong, then,
for doctors sometimes to act against a duty to preserve life in order to relieve
pain, just as they could sometimes act against a duty not to intend pain in
order to save a life?
To summarize, I have constructed a three-step argument for physician-assisted
suicide and euthanasia. Assuming patient consent:
1. We may permissibly cause death as a side effect if it relieves pain, because
sometimes death is a lesser evil and pain relief a greater good.
2. We may permissibly intend other lesser evils to the patient, for the sake
of her greater good.
3. Therefore, when death is a lesser evil, it is sometimes permissible for
us to intend death in order to stop pain.
Thus, suppose we accept that it is sometimes permissible to knowingly
shorten a life by giving pain-relieving medication, and agree, too, that it
is sometimes permissible for a doctor to intend a lesser evil in order
to produce a greater good. How, then, can it be wrong to intentionally
shorten a life when that will produce the greater good?2
I don't expect that everyone will immediately find this argument compelling.
I suspect that many--including some who are inclined to agree with the conclusion--will
feel that death is different, so to speak. While they agree that we may intend
pain, if it is a lesser evil, in order to save a life, they think it is impermissible
to intentionally hasten death in order to relieve pain. I will address this
concern later. But first I want to add another set of considerations that
support euthanasia and physician-assisted suicide.
III. An Argument for Duty
According to the three-step argument, a doctor is permitted to give
morphine for pain relief, even though he knows it will expedite the patient's
death, if death is the lesser evil. But I think we can say more. Suppose,
as I have stipulated, that giving morphine is the only way for a doctor to
relieve a patient's suffering. A doctor, I assume, has a duty to relieve suffering.
I conclude that the doctor has a duty to relieve suffering by giving
the morphine, if the patient requests this. He cannot refuse to give the morphine
on the ground that he will be a killer if he does.
If doctors have a duty to relieve pain, and even being a killer does not
override this duty when the patient requests morphine for pain relief, then
perhaps they also have a duty, not merely a permission, to kill their patients,
or aid in their being killed, intending their deaths in order to relieve suffering.
Now we have a new argument. Assuming patient consent:
1. There is a duty to treat pain even if it foreseeably makes one a killer,
when death is the lesser evil and no pain is the greater good.
2. There is a duty to intend the other lesser evils (e.g., amputation) for
a patient's own greater good.
3. There is a duty to kill the patient, or assist in his being killed, intending
his death when this is the lesser evil and pain relief the greater good.
I think this argument, too, is compelling, but will concentrate here on the
case for permissibility.
IV. Is Killing Special?
As I indicated earlier, a natural rejoinder to the three-step argument for
euthanasia and physician-assisted suicide is to emphasize that "death is different."
But how precisely is it different, and why is the difference morally important?
Perhaps it will be said, simply, that the doctor who intends the death of
his patient is killing. Even if intending a lesser evil for a greater
good is often permissible, it might be condemned when it involves killing.
Killing, it might be said, is not on a par with other lesser evils.
But this does not suffice to upset the three-step argument. For giving a
lethal injection of morphine to relieve pain also involves killing, and we
approve of giving the morphine. To be sure, a patient's right to life includes
a right not to be killed. But that right gives us a protected option whether
to live or die, an option with which others cannot legitimately interfere;
it does not give one a duty to live. If a patient decides to die, he is waiving
his right to live. By waiving his right, he releases others (perhaps a specific
other person) from a duty not to kill him, at least insofar as that duty stems
from his right to live.3 The duty not to kill him may also stem from their duty not
to harm him, even if he so wishes; but I have stipulated that the doctor is
to kill only when death is the lesser evil.
A more compelling version of the objection is, however, waiting in the wings.
This one points not merely to the fact of killing, but to intentional killing.
It claims that there is something distinctive about intending death, and that
this distinction makes a large moral difference. In particular, acting with
the intention to bring about death as a lesser evil requires that we treat
ourselves or other persons as available to be used for achieving certain goods--in
particular, the reduction of suffering. In euthanasia and physician-assisted
suicide, we intentionally terminate a being with a rational nature--a being
that judges, aims at goals, and evaluates how to act.4
We have no such intention to use a person as a mere means when we aim at such
lesser evils as destruction of a leg. Indeed, one of the things that seems
odd about killing someone only if he is capable of voluntarily deciding in
a reasonable way to end his life is that one is thereby ensuring that what
is destroyed is a reasoning, thinking being, and therefore a being of great
worth. This will not be so if the person is unconscious or vegetative or otherwise
no longer functioning as a rational being. Obviously, people take control
of their lives and devote their rational natures to the pursuit of certain
goals within those lives; but, it is claimed, when this is appropriate, they
do not aim to interfere with or destroy their personhood but set it in one
direction or another.
The idea that there are limits on what we may do to ourselves as persons
derives from Immanuel Kant. In his moral writings, Kant said that rational
humanity, as embodied in ourselves and others, is--and should be treated as--an
end in itself, and not a mere means to happiness or other goals. The fact
that one is a judging, aiming, evaluating rational agent has worth in itself.
To have this value as a person is more like an honor to us (Kant called it
"dignity") than a benefit that answers to some interest of ours. Thus my life
may have worth, even if my life is not a benefit to me (and my death would
benefit me) because goods other than being a person are outweighed
by bads. The worth of my life is not measured solely by its worth to me in
satisfying my desires, or its worth to others in satisfying theirs. According
to Kant, then, it is wrong for others to treat me as a mere means for their
ends, but equally wrong for me to treat myself as a mere means for my own
ends: As others should respect my dignity as a person by not using me merely
as a means for their purposes, I should have proper regard for my own dignity
as a person, and not simply use myself as a means for my own purposes. But
that is precisely what I do when I aim at my own death as a way to eliminate
pain. So I ought not to pursue that aim, and therefore ought not to consent
to a morphine injection aiming at death, or give one to a patient who has
Before assessing this Kantian argument, I want to justify focusing it on
intentional killing rather than other ways of intentionally contributing to
a death. Consider a patient who intends his own death and therefore wants
life support of any sort removed. Suppose, for the sake of argument, that
we disapprove of this intention. Suppose, too, that we disapprove of a doctor's
agreeing to remove treatment because he also intends this patient's death.
But while we may disapprove of the intentions and conduct, acting on that
disapproval would require us to force life support on the patient,
and he has a right that we not do this. Our opposition to his intentions and
the doctor's is trumped by our opposition to forced invasion of the patient.
So we permit the patient and doctor to act--to remove treatment--intending
death. Consider, in contrast, a patient who intends his own death and, therefore,
requests a lethal injection or pills. Suppose, once more, that we disapprove
of this intention. Acting on our opposition would require us to refrain from
invading him with a lethal injection or refuse the pills. But it seems clear
that the right not to be invaded with treatment against one's will is stronger
than the right simply to be invaded (with a lethal injection) or given pills.
So the fact that we must terminate treatment, even when the patient and doctor
intend the patient's death, does not show that it is permissible to kill the
patient or assist him in killing when he and his doctor intend his death.
Correspondingly, an objection to intentional killing need not imply an objection
to terminating treatment for someone who intends his own death.5
I turn now to the Kantian-style argument against aiming at one's death (or
aiming at another's death with his consent). In assessing this argument, we
must distinguish three different ways in which one may treat a person as a
1. Calculating the worth of living on in a way which gives insufficient weight
to the worth of being a person.
2. Treating the nonexistence of persons as a means to a goal (e.g., no pain).
3. Using persons in order to bring about their own end.
The first idea is that being a person has worth in itself and is not merely
a means to an overall balance of other goods over evils in the person's life.
On this interpretation, we treat persons as a mere means if we give inadequate
weight in our decisions to the value of our existence as persons; if we do,
then death may seem a lesser evil. But even when there are few goods in life
besides the capacity to be a rational agent, the loss of life--and therefore
the loss of that capacity--may still be a greater evil than pain.
Though I do not doubt that this idea has force, it can equally well be given
as a reason for not terminating a course of treatment, even when one merely
foresees one's death. Because this way of treating a person as a mere means
does not distinguish the morality of intending death from the morality
of merely foreseeing death, it cannot be used to explain why intentional
killing in particular is impermissible.6
It might be said that this observation should prompt us to rethink the permissibility
of killing in, for example, the case of morphine for pain relief, where death
is foreseeable but not intended. For it might be said that the concern to
treat people as ends and not mere means can be met only by giving overriding
weight to the value of rational humanity, and that this requires that we refrain
from acting in ways that we foresee as leading to death, and not simply from
intentional killing. This response seems unjustified. Suppose life involves
such unbearable pain that one's whole life is focused on that pain. In such
circumstances, one could, I believe, decline the honor of being a person.
In so doing, we need not treat our life as mere means to a balance of good
over evil. We might acknowledge the great (and normally overriding) value
of being a person, and believe it is right to go on in life even if it has
more pain than other goods besides rational agency: though we reject the thought
that rational agency is merely a means to happiness, we allow that some bad
conditions may overshadow its very great value.
What, then, about the second and third interpretations of the idea of using
persons as mere means? To see the difference between them, consider an analogy:
My radio is a device for getting good sounds and filtering out bad sounds.
It is a means to a balance of good sounds over bad ones. Suppose it stops
performing well, that it only produces static, but cannot be turned off. I
can wait until its batteries run down and not replace them, or I can smash
it now, thus using the radio itself to stop the noise it produces. Either
way, I would see its death as I saw its life, as a means to a better balance
of good over bad sounds. While I have always seen my radio as a mere means
to an end, if I smash it, I use it as a means to its end (termination): This
is sense (3).7 If I let the radio run down, intending
its demise, but do not smash it--I see it wasting away and do not replace
its parts--then I do not see it as a means to its own end, but I do see its
end as a means to a better balance of sounds. This is sense (2).
Active suicide is analogous to smashing the radio: the person uses himself
as a means to his own death. Some people find this complete taking control
of one's life particularly morally inappropriate, perhaps because they think
our bodies belong to God and that we have no right to achieve the goal of
our own death by manipulating a "tool" that is not ours (or intending that
others manipulate it). This objection is not present if--here we have sense
(2)--we terminate medical assistance with the intention that the system run
down, aiming at its death. For then we achieve the goal of death by interfering
with what is ours (the medication), not God's. Here we have another reason
why someone may object to killing but not to terminating treatment, even if
accompanied by the intention that the system run down: unlike intentional
killing, terminating treatment does not involve using persons to bring about
their own end. Some say, though, that this way of using persons as means is
also more objectionable than merely foreseeing the death. They say that if
we terminate medical assistance, intending death, we do not merely treat our
life as a means to greater good over bad, but treat our death (the end
of our life) as a means to greater good over bad.
How much weight, then, should be placed on the second and third senses of
"using a person as a means"? Should they really stand in our way? I believe
not. It cannot be argued, at least in secular moral terms, that one's body
belongs to someone else and that one cannot, therefore, use it as a means
to achieve death. Notice also that if your body belonged to someone else,
it isn't clear why you should be permitted to use it by administering morphine
to stop your pain when you merely foresee tht this will destroy the body.
We aren't usually permitted to treat other people's property in this way either.
Nor does it seem that treating one's death as available for one's purposes
(i.e., being rid of pain) is necessarily a morally inappropriate attitude
to take to oneself--so long as there is no failure to properly value the importance
of just being a person. If this is right, then, at least sometimes, a patient
would do no wrong in intending or causing his death. At least sometimes, a
doctor who helped him by giving pills would also do no wrong merely because
he killed, or assisted killing, aiming at death.
The strongest case for such conduct can be made, I believe, if the overriding
aim is to end physical pain. The need to do this may be rare with modern techniques
of pain control, but still the patient has a "disjunctive" right: either to
adequate pain control or the assistance in suicide of a willing doctor. Psychological
suffering which is a reaction to one's knowledge or beliefs about a state
of affairs is a weaker case. The test I suggest here is: Would we give a drug
to treat psychological suffering if we foresaw that it would rapidly
kill as a side effect? If not, then giving pills to a patient intending that
they kill him in order to end psychological suffering would not be permissible.
This same test can be applied to other reasons that might be offered for seeking
death by euthanasia or physician-assisted suicide. For example, would we allow
a patient to use a drug that will rapidly cause death (rather than a safer
one) if it will save him money? If not, then we may not perform euthanasia
or physician-assisted suicide to stop the drain on his family finances. Would
we give a demented patient a drug that unraveled the tangled neurons that
caused his dementia but which we foresaw would rapidly kill him as a side
effect? If not, then why should we be permitted to give him pills, intending
his death? Of course, the application of this test may yield positive responses
rather than negative ones to these questions.
V. A Philosopher's Brief
I have argued that if it is permissible to treat someone in his best interests
though we foresee that this treatment will rapidly cause death, it
is permissible to kill or assist in killing someone intending his death
when this is in his best interest. (Consent is required in both cases.) In
their recent "philosophers' brief," Ronald Dworkin, Thomas Nagel, Robert Nozick,
John Rawls, Thomas Scanlon, and Judith Jarvis Thomson8 also reject a blanket ban on physician-assisted suicide, but their
strategy of argument (and some of their conclusions) differ from mine. They
argue from the premise that it is permissible to omit or terminate treatment
with the intention that the patient die to the conclusion that it is permissible
to assist in the killing of the patient intending death. I think this approach
neglects the moral force of the distinction between killing a person and letting
Dworkin et al. wrote as amici curiae in the "assisted suicide cases"
that the Supreme Court heard this past January.9
The cases came on appeal from two circuit courts, which had both ruled that
government cannot simply prohibit doctors from prescribing medication that
would hasten the death of patients who request such medication. The philosophers
urge the Supreme Court to uphold those decisions. One part of their argument
builds on the Court's 1990 decision in Cruzan v. Missouri, in which
the Court majority assumed (if only for the sake of argument) that patients
have a constitutional right to refuse life-preserving treatment.10
According to Dworkin et al., the existence of a right to refuse treatment
also implies a right to assistance in suicide: if, as Cruzan indicates,
it is permissible for doctors to let a patient die (even when patient and
doctor intend the patient's death), then it is permissible for doctors to
assist in killing. In a preface to the philosophers' brief written after the
Supreme Court heard oral argument on the case, Ronald Dworkin notes that several
justices rejected this link between Cruzan and the assisted suicide
cases. They sought to distinguish them by reference to a "common-sense distinction"
between the moral significance of acts and omissions: assisting suicide is
an act, and thus requires a compelling moral justification; in contrast, not
providing treatment is an omission, a matter of "letting nature take its course,"
and can be justified more easily. Dworkin says that "the brief insists that
such suggestions wholly misunderstand the `common-sense' distinction, which
is not between acts and omissions, but between acts or omissions that are
designed to cause death and those that are not."
I agree that the act/omission distinction will not bear much moral
weight in this setting: When a doctor removes treatment by pulling a plug,
she acts, though she does not necessarily kill. If the doctor is terminating
aid she (or an organization she is part of) started, then I think she does
not kill the patient but lets him die. Consider an analogy: I am saving someone
from drowning and decide to stop. Even if I have to push a button to make
myself stop, I do not kill the person but let him die. When the doctor gives
morphine to ease pain, foreseeing it also causes death, she also acts and
even kills. But I part company with these philosophers when they argue that,
once a patient has consented, we can always move from the permissibility of
removing treatment while intending the patient's death to the permissibility
of assisted suicide. Killing is not on a moral par with letting die. Let me
explain by reference to some cases.
In the first, doctors act against their patients' wishes to live.
Dworkin et al. agree that a doctor may permissibly deny an organ to a patient
in order to give it to another, but not kill a patient to get his organ
for another. This is not, they say, because of a moral difference between
letting a patient die and killing him, but because the doctor merely foresees
death in the first case, whereas he intends it in the second. If this showed,
however, that killing made no moral difference, it would imply that it is
permissible for a doctor, in order to transplant an organ into a patient,
to use a chemical that he foresees will seep into the next room where
another patient lies, killing that patient. For in this case, the doctor does
not intend to kill the patient in the other room, but only foresees his death
as a side-effect of the chemical. Presumably, though, transplanting with this
effect is wrong because it is a killing, albeit foreseeable rather than intended.
So in cases in which we merely foresee death, killing may be wrong, even if
letting die is not.
Killing can also be a significant factor in cases where a patient does not
want to die because letting die with the intention that death occur might
be permissible even if killing with such intention is not. It is true, as
the brief says, that a doctor who lets his patient die of infection against
the patient's will, intending that he die so that his organs are available,
has done something wrong, as has a doctor who kills the same sort of patient,
intending him to die. In both cases, the doctors aim against the welfare of
their patients and violate their rights. The first doctor violates a right
to treatment; the second doctor violates a right against being killed. But
this does not always imply, as Dworkin et al. think, that "a doctor violates
his patient's rights whether the doctor acts or refrains from acting against
the patient's wishes in a way that is designed to cause death." For example,
suppose that someone does not want to die but it would be in his interest
to die. If a treatment is experimental, or in general something to which the
patient has no positive right, it may be permissible to deny it to him because
death would be in his own best interest. I do not believe the patient acquires
a right to have the experimental therapy merely because the doctor's reason
for refusing it is that he aims at the patient's death. Still, it would be
morally wrong to kill the patient if he did not want to die, even if it were
in his interest to die.
Consider next the type of case in which the patient wants death. Here, the
distinction between killing and letting die makes a moral difference when
deciding on the scope of permissible refusal of treatment versus permissible
assistance to killing. Dworkin et al. seem to suggest the scope should be
the same, saying that if a doctor can turn off a respirator, he can prescribe
pills. But a mentally competent patient may legally refuse treatment, intending
to die, even when it is not in his best interest to do so and, on many occasions,
even when he could be cured. Presumably, in many of these cases, he could
also insist on terminating treatment, even if his intention is to die. Furthermore,
even if the doctor in these cases improperly intends that the patient die,
the treatment must be terminated. This is because the alternative to letting
the patient die is forcing treatment on him. We think a competent patient's
right not to be physically invaded against his will is typically stronger
than our interest in his well-being (though this right is, to be sure, not
absolute and can sometimes be overridden by considerations of public safety).
But if he asks for assistance in killing himself when it is not in his interest
to be killed, it might well be morally impermissible to kill him. Contrary
to what Dworkin et al. say, therefore, a doctor might in some cases be permitted
and even required to turn off a respirator but not permitted to give pills.11
The alternative to letting die, then, has such a morally objectionable feature--forcing
treatment which he has a right we not do--that even if we think the patient's
and doctor's goals are wrong, we must terminate aid. In contrast, the alternative
to assisted suicide may simply be leaving the patient alone; this often does
not violate any of his rights against us, and so we can, and sometimes we
should be required to, refuse to help because we disapprove of his goals.
Many people--including the Supreme Court justices Dworkin et al. cite--might,
then, reasonably distinguish refusing treatment and thus letting a patient
die from assisting in a suicide. The move from Cruzan's right to refuse
treatment to the permissibility of assisted suicide is, therefore, not generally
available. Still, my discussion here indicates that we may have other reasons
to accept the fundamental moral conclusion: that assisted suicide (and euthanasia)
are sometimes morally permissible.12
* An earlier version of this paper was given as a talk
at the Plenary Session of the American Academy of Forensic Sciences in New
York, in February 1997.
1 Only "may," because some doctors aim simply to give patients
a choice whether to live or die.
2 I first presented this argument in Creation and Abortion
(New York: Oxford University Press, 1992), pp. 33-35, and again in Morality,
Mortality, Vol. II (New York: Oxford University Press, 1996), pp 194-98.
3 Notice that this waiver seems to be morally necessary
even when the doctor wishes to give morphine that will kill as a foreseen
side effect. This means doctors should get permission for giving the morphine
for pain relief as well as for giving it to deliberately kill. (I do not believe
they always do so.)
4 We also terminate human life considered independently
of whether it is the life of a rational being. It may seem harder to justify
destroying a person than a human life that lacks qualities required for personhood--for
example, a functioning brain. But I will assume that one could substitute
"human life" for "person" in the argument I give against intentional killing
and in my response to that argument.
5 In contrast, suppose that a patient who intends his own
death is also suffering great pain that only morphine will stop. He asks for
the morphine, not because it will stop the pain, but because he knows it will
kill him. If it would not kill him, he would not ask for it. Does he have
a right that the doctor give him the morphine? If he does, then the doctor
is not at liberty to refuse simply because of the patient's intention,
any more than he could refuse to terminate treatment because of the patient's
intention. Indeed we might not be permitted to interfere with the doctor's
giving morphine in this case even if he gave it only because he intended
death. I owe this case to Timothy Hall.
6 Kant thought we had a duty to actively preserve rational
humanity and hence we should not too lightly do what we foresee will lead
to its end. Still, he allows that we may sometimes engage in conduct though
we foresee it will result in our deaths, but we may never aim at our deaths.
7 If I see someone else destroy it and do not interfere,
I may be intending its use as a means to its own end, though I do not myself
8 New York Review of Books, March 27, 1997.
9 State of Washington et al. v. Glucksberg et al. and
Vacco et al. v. Quill et al.
10 Cruzan v. Missouri, 497 U.S. 261 (1990). The Court
did not, however, agree that Nancy Cruzan had clearly decided, before an automobile
accident that left her in a persistent vegetative state, that she would refuse
11 It seems to me that the lower court in the Washington
case also overlooked this point when they said that whether we assist killing
or let die is not morally or legally crucial. For even that court was concerned
to limit the doctor's right to assist in killing the patient to cases where
the patient's life is going to end shortly anyway and when death is not against
his interests. But the right to refuse treatment and have it terminated applies
much more broadly. If, however, the distinction between killing and letting
die as such makes no moral or legal difference, then refusing treatment should
be permitted no more broadly than assisting killing. See Compassion in
Dying v. Washington 79 F.3d 790 (9th Cir. 1996).
12 I am grateful to Joshua Cohen, Timothy Hall, and Seana
Shiffrin for their comments.