Almost 30 years ago, James Rachels (writing in what must be one of the most well known papers in medical ethics) described what he took to be the prevailing view of euthanasia as follows¹:

The distinction between active and passive euthanasia is thought to be crucial for medical ethics. The idea is that it is permissible, at least in some cases, to withhold treatment and allow a patient to die, but it is never permissible to take any direct action designed to kill the patient.

In the subsequent ethics literature on euthanasia, there has been a widely accepted euthanasia taxonomy comprising two key distinctions. Firstly, there is Rachels’ distinction between euthanasia performed by killing the patient (active euthanasia) and euthanasia performed by omitting to prolong the patient’s life (passive euthanasia). And second, cutting across this active–passive distinction, is a distinction between voluntary, non-voluntary, and involuntary euthanasia, depending on whether patients autonomously request their death, are unable competently to give consent, or are competent but have their views on the matter disregarded (or overruled).

In general this categorisation is useful, since it reveals the possibility of finding some types of euthanasia less morally objectionable than others. So although of course the taxonomy per se cannot answer any moral questions, it does seem to provide a useful framework within which to think about them. In recent years, however, this standard categorisation and in particular the idea of passive euthanasia have been subjected to attacks from a number of authoritative sources. For example, the House of Lords Select Committee on Medical Ethics describes “the term passive euthanasia” as “misleading”² and the British Medical Association calls the expressions “active” and “passive” euthanasia “ambiguous and unhelpful”, claiming that³:

Confusion may … arise when the withdrawing or withholding of life-prolonging treatment which is not providing a benefit to the patient is described as “passive euthanasia”.

This attack on the very idea of passive euthanasia is expressed in a particularly clear and explicit way by an “ethics task force” established by the European Association of Palliative Care (EAPC) in February 2001.⁴ The EAPC Task Force rejects the standard taxonomy in two ways. Firstly, it claims that the distinction between active and passive euthanasia is in some way inappropriate: euthanasia is active by definition and hence “‘passive’ euthanasia is a contradiction in terms—in other words, there can be no such thing”.⁵ Secondly, it asserts that there can be no such thing as non-voluntary or involuntary euthanasia, since “medicalised killing of a person without the person’s consent … is not euthanasia: it is murder”. So “non-voluntary euthanasia” is a contradiction in terms too. In this paper, we focus exclusively on the first claim, the one concerning passive euthanasia—not least because this view is becoming popular amongst palliative care practitioners and therefore stands in need of careful critical assessment. That it is gaining support is evidenced both by our numerous conversations with healthcare professionals and by comments on the EAPC Ethics Task Force paper submitted to the journal Palliative Medicine, whose editors tell us that “the suggestion that the differentiation between active and passive euthanasia is inappropriate” was “widely welcomed”.⁶

The Task Force’s paper does not itself contain any arguments for the claim that passive euthanasia is a contradiction in terms, and hence in what follows we will construct for ourselves what seems to be the best case in support of this view before subjecting it to critical assessment. We do not therefore suggest that the members of the Ethics Task Force actually subscribe to all of these arguments, merely that these are the best possible arguments for the Task Force’s view.

WHAT IS PASSIVE EUTHANASIA?

Before considering the arguments, we need to clarify just what passive euthanasia is taken to be by those who do find it an acceptable category. Like all forms of euthanasia, it involves the intention to hasten death in the patient’s interests (because of their expected negative quality of life). What is standardly taken to mark off passive as opposed to active euthanasia is that the former hastens death by not providing something which would, if provided, delay death—that is, passive euthanasia involves withdrawing or withholding life-prolonging medical treatment. So there are (at least) three necessary conditions for the occurrence of passive euthanasia:

1. there is a withdrawing or withholding of life-prolonging treatment

2. the main purpose (or one of the main purposes) of this withdrawing or withholding is to bring about (or “hasten”) the patient’s death

3. the reason for “hastening” death is that dying (or dying sooner rather than later) is in the patient’s own best interests.

It is important to note that not all cases of withdrawing or withholding life-prolonging treatment are cases of passive euthanasia. The grounds for passive euthanasia are, as we have seen, the interests of patients, where their expected quality of life is so poor that life will be worse for them than death. But there are many other reasons for withdrawing or withholding treatment. Firstly, treatment might simply be futile and hence incapable of benefiting the patient. Secondly, the treatment may not be cost-effective. Whether or not it can be justified, refusal to treat on grounds of cost-ineffectiveness does not amount to passive euthanasia, since it does not meet condition 3 (which holds for any form of euthanasia), namely that death is being hastened in the patient’s own interests. Thirdly, treatment may be withheld or withdrawn because it is excessively burdensome or harmful. This fails to amount to passive euthanasia because it breaches condition 2: the health carer’s intention is not to bring about death, but to protect the patient from a burden or harm. In such cases, whether the healthcare professional’s intention is to bring about death can be established by using the following counterfactual test: if the patient does not die, has the health carer succeeded in his or her aim? In the case of passive euthanasia, the answer will be “no” because the health carer was aiming at the patient’s death. But in the case of withdrawing or withholding treatment because it would be burdensome or harmful, then the health carer can have succeeded in this aim—to protect the patient from a particular burden or harm—even if the patient pulls through without treatment. A fourth possible reason for withdrawing or withholding treatment is that the patient has refused the treatment in question. Any competent adult patient has the right to do this, and a healthcare professional’s conforming to this refusal does not amount to passive euthanasia, since again it fails to meet condition 2. Even if the refusing patient intends to hasten their own death, it does not follow that the healthcare professional shares their intention. The health carer’s aim may be purely to respect the wishes and the autonomy of a competent patient; they need not share the patient’s death directed intentions at all, and again might be very glad if the patient survives in spite of refusing treatment.

With the definition of passive euthanasia established, we can now turn to a closer examination of the EAPC Ethics Task Force’s views. There seem to be three main arguments for its view of passive euthanasia, and we shall assess each in turn.

THE WRONGNESS OF EUTHANASIA ARGUMENT

According to the first argument, euthanasia is always morally wrong. But the behaviour which is described in the standard taxonomy as “passive euthanasia” is not morally wrong. Therefore, it cannot really be a form of euthanasia and so there is no such thing as passive euthanasia.

Put like this, the argument seems simplistic and implausible. However, there is perhaps a slightly better way of articulating it. Some concepts, so the argument goes, are moral concepts, or at least have a moral dimension. For example, “murder” (understood here as a moral rather than a legal term) is typically defined as wrongful killing. Or, to put it another way, one necessary condition for an act being a murder is that it is wrong. So perhaps euthanasia is like murder and part of its definition is that it is a wrongful act.

There are, however, a number of objections to this argument. To being with, it assumes too much. It starts from the claim that euthanasia is never justified, and builds this into the definition of euthanasia, excluding on grounds of conceptual incoherence the possibility of there being a permissible form of euthanasia. This not only excludes the possibility of passive euthanasia (assuming this to be morally permissible) but also excludes the possibility of a moral debate about the ethical standing of euthanasia. And surely we should not foreclose on moral debate in this way. It would be better if our definitions at least left open the possibility that active euthanasia is (sometimes) justified, and that passive euthanasia is (sometimes) wrong.

This point is further strengthened if we consider two “real life” counterexamples. Firstly, of course, there are lots of people who think that active voluntary euthanasia is (sometimes) permissible. This moral belief may or may not be true, but are we really supposed to believe that these people are believers in something self-contradictory or incoherent? Also, what if we were to discover somehow that the view of the “pro” active voluntary euthanasia organisations was the correct one and that active euthanasia is in fact morally permissible. Ought we to conclude from this that there are not really any cases of euthanasia, because all of the apparent cases have turned out not to be wrong, and hence not euthanasia? Surely not.

Conversely, some moral conservatives believe that even passive euthanasia is wrong. For example, Anne Winterton MP tried (unsuccessfully) to ban passive euthanasia by introducing a private members bill in the UK parliament. The bill said⁷:

It shall be unlawful for any person responsible for the care of a patient to withdraw or withhold from the patient medical treatment … if his purpose or one of his purposes in doing so is to hasten or otherwise cause the death of the patient.

Presumably we should not rule out completely (and a fortiori not rule out by definition) the possibility that Anne Winterton and others are correct about the wrongness of passive euthanasia. But, if combined with the wrongness of euthanasia argument, what is entailed is that if Winterton et al were somehow shown to be correct, then passive euthanasia could exist after all (since it would, like active euthanasia, be wrong). So the argument is, to say the least, strange insofar as it makes the very existence of euthanasia (both active and passive) dependent on its moral status.

A fundamental problem with the wrongness of euthanasia argument then is that the evaluation is driving the conceptualisation. It is hard to see why we should give our evaluations priority in this way, and furthermore this is something which the EAPC Ethics Task Force explicitly denies doing. It claims that its definition of euthanasia “say[s] nothing about the norms and values associated with what is defined”, and that “whether or not euthanasia may be justified … is another matter … A sharp distinction, therefore, exists here between what ‘is’ and what ‘ought’ to be.”⁸ The is–ought distinction which the Task Force endorses here undermines the wrongness of euthanasia argument.

THE CAUSATION ARGUMENT

The second argument focuses on problems with causation. Uncontentiously, euthanasia causes death. But, according to the causation argument, omissions cannot cause anything. Passive euthanasia (so-called) is an omission. So passive euthanasia cannot cause death and hence cannot really be euthanasia.

The causation argument is also seriously flawed. Most importantly, the general claim at the heart of the argument—that omissions cannot be causes—does not appear to be true. What caused the crash at the junction? One of the drivers did not signal before turning right—he omitted to signal his intentions. What caused the clever student to fail her exams? She did not bother to revise for them—she omitted to do any work. We routinely attribute causal explanatory power to a wide variety of omissions. This objection alone seriously erodes the causation argument. But it is not alone; it goes hand in hand with an even stronger objection in the same style.

There are some circumstances in which omissions can clearly kill. Suppose a patient has a raging infection which the doctor can cure simply by administering antibiotics. But he negligently or malevolently does not prescribe them—he omits to provide the appropriate treatment, thereby killing the patient. Suppose the company refuses to provide safety devices for workers carrying out high risk activities—the boss thinks they cost too much, the money can be more profitably spent elsewhere, perhaps there will not be an accident, at least not this year. And then a worker dies, when if he had been issued the standard (though expensive) safety kit he would have lived. The failure to provide him with these devices killed him. Suppose that a child is drowning in shallow water and that a passing adult, who is a strong swimmer, decides not to save her—perhaps because he hates her parents, or because he is on his way to a football match and does not want to be late. The passer-by decides not to intervene when it would have been easy for him to do so and his non-intervention kills the child. Instances such as these can be multiplied endlessly: omissions can kill. But in that case, passive euthanasia can kill even if, unlike the examples given above, it is morally permissible.

In response, it might be argued that the moral status of an omission can affect whether it counts as a cause or not (and hence whether it counts as a killing) and in particular that only impermissible omissions can be causes. For example, when parents wrongfully fail to care for their children, or pet owners negligently fail to feed their animals, we naturally regard these omissions as causes of death (when death occurs). The main reason for regarding them as causally efficacious is that that they are breaches of a positive moral duty to act; the people in question are not doing what is expected of them. Crucially, this is what differentiates parents’ and pet owners’ non-actions from the causally irrelevant non-actions of unconnected others. For it is equally true of both the neglectful parent and of the stranger who lives on the other side of town that but for their not feeding the child, the child would have survived—and so any “causal” difference between them, it could be argued, must be based on the fact that one has a positive duty, while the other (we suppose) does not.

But our intuitions here are not very stable. Suppose 10 patients are hooked-up to apparatus which is delivering life-prolonging treatment. For seven of the patients, the treatment is in their interests, and should be maintained; for the remaining three, it is not, and in those cases, it would be morally permissible to withdraw the treatment. One switch controls all the machinery and someone deliberately throws the switch. All 10 patients die. Is it plausible to say that in seven of the cases (the seven for whom it would be morally impermissible to withdraw the treatment) throwing the switch was the cause of death, and hence it killed the patients, whereas in the three cases where it was in the patients’ interests for the treatment to be withdrawn, throwing the switch was not the cause of death? We are more likely to say that if some of these patients were killed by the withdrawal of treatment, then all of them were. So we cannot securely appeal to intuitions about such cases to enable us to assert that permissible omissions can never amount to killing. (It might be objected here that, in our example, throwing the switch is an action. However, this example would work equally well if failing to throw the switch was used instead as the putative cause of death.)

Before moving onto the third argument for the EAPC Task Force view, we should briefly mention a supplementary argument that is sometimes used to back up the causation argument. This asserts that in cases of (so-called) passive euthanasia, it is the disease itself which causes death, rather than the withholding or withdrawing of treatment. So, in the context of terminal illness, failing to delay the death need not amount to causing it. This style of argument interestingly mirrors well known remarks by Devlin J (later Lord Devlin) in Adams (a significant English legal case from 1957)⁹:

“Cause” means nothing philosophical or technical or scientific. It means what you twelve men and women sitting as a jury would regard in a commonsense way as the cause. Manifestly there must be cases in hospitals that are going on day after day in which what a doctor does by way of giving certain treatment prolongs or shortens life by hours or even longer. The doctor who decides to administer or not to administer a drug is not, of course, thinking in terms of hours or minutes of life. He could not do his job properly if he were. If, for example, because a doctor had done something or has omitted to do something, death occurs at eleven o’clock instead of twelve o’clock, or even Monday instead of Tuesday, no people of common sense would say “Oh, the doctor caused her death”. They would say that the cause of death was the illness or the injury, or whatever it was, which brought her into the hospital, and the proper medical treatment that is administered and that has an incidental effect on determining the exact moment of death is not the cause of death in any sensible use of the term.

The thought here seems to be that the real or underlying or fundamental cause of death is the disease (not the withdrawing or withholding of treatment) and this idea is (supposedly) bolstered by the fact that, were it not for the disease, the patient would not have presented to the healthcare professional in the first place and so questions about withdrawing or withholding treatment would not even have arisen.

This supplementary argument though is also flawed. For as Devlin rightly suggests, if this argument works at all then it “proves too much”, since it applies both to acts and omissions, both to active and passive euthanasia. For one might in the same way argue that, even in cases of active euthanasia, the underlying or fundamental cause of death is the disease (and on the same grounds—that were it not for the disease, the patient would not have presented to the healthcare professional in the first place). So either we must reject this supplementary argument altogether or we must accept it and also accept (at least for patients who are terminally ill) that neither (so-called) active euthanasia nor (so-called) passive euthanasia causes death and therefore that neither practice is really euthanasia. And whichever option we choose, the view that only active euthanasia causes death (and hence that all euthanasia is active) would remain unsupported.

THE BAD CONSEQUENCES ARGUMENT

There are two versions of the third argument for the EAPC Task Force view, both of which appeal to the alleged bad consequences of accepting the category of passive euthanasia.

According to the first version, the behaviour which the standard taxonomy categorises as passive euthanasia is widely, and rightly, regarded as morally acceptable. However, if we call this passive euthanasia, then people will be encouraged to think that all kinds of euthanasia (including the active form) are acceptable. And since active euthanasia is wrong, it would be very bad if people regarded it as permissible. This is a slippery slope argument, taking us from the first (apparently permissible) step of calling certain omissions “passive euthanasia”, to the final position at the bottom of the slope in which we tolerate active euthanasia, encourage people to seek it out, and perhaps even put pressure on those who do not wish to die to accept euthanasia.

The second version of this argument also raises a worry about the consequences of acknowledging a category of passive euthanasia. But here the worry is not about a slide towards murder, but rather about the erosion of trust between patient and doctor. According to this version, patients are not normally very good at making fine distinctions such as that between active and passive euthanasia. So if we tell them that passive euthanasia is acceptable and is being practised then they are likely to get confused and think that active euthanasia is being practised as well. This will damage attitudes towards the medical profession (especially towards those in palliative care). People will fear that health carers are no longer committed to preserving lives and may constitute a threat to their patients. This undermining of trust will deeply damage patient care. Patients will no longer feel free to tell their doctors everything about their physical and emotional condition, for fear of evincing unwanted suggestions about euthanasia. In the light of this risk, so the argument goes, we ought not to call passive euthanasia “passive euthanasia”, to avoid triggering this damaging reaction. Even better, we should adopt official definitions of euthanasia which do not classify (so-called) passive euthanasia as euthanasia at all.

Although these arguments are different, in that they foresee different bad consequences, nonetheless their structural similarities mean that they are vulnerable to the same kind of objections.

The first objection is that the arguments (again) assume too much. They simply presuppose that active euthanasia is wrong and that passive euthanasia is permissible, and it would be much better if arguments could be found which did not rely on such controversial assumptions. For many people when presented with the slippery slope argument outlined above will simply say that what’s at the bottom of the slope (the acceptance of active euthanasia) is not in fact a bad consequence, but rather a good one.

A second objection appeals to a quite general methodological problem with this argument: that it tries to settle an ontological issue (a question about what kinds of things exist) by citing (expected) good or bad outcomes. But we cannot establish what kinds of things or events there are in the world by considering the good or bad consequences of people believing (or not) in their existence. We cannot, for example, determine whether there really are such things as bacteria, or genes, or God, or tooth fairies by working out the likely consequences of popular acceptance of these entities, even though such acceptance may well have important social implications. Perhaps a pertinent example of this is the concept of “disorder”.¹⁰ If we classify a particular state as a disorder (for example, chronic fatigue syndrome (CFS)) this may well have significant consequences for the bearers of that state. Some of these may be good (for example, more funding for research and treatment) and some bad (for example, possible long term stigmatisation or overdiagnosis). But weighing up the positive and negative consequences of classifying CFS as a disorder will not settle whether or not CFS actually is a disorder. Indeed, such consequentialistic calculations are entirely irrelevant if our concern is the actual disease status (or otherwise) of CFS. For if the consequences of classifying a condition as a disorder are genuinely bad or good, the most that can do is give us a reason for keeping quiet about (or exaggerating) the real status of the condition. The bad consequences cannot stop a condition from being a disorder. Similarly we cannot settle the ontological question of whether there is such a thing as passive euthanasia by appealing to the alleged bad consequences of answering that question in the affirmative.

So even if we allow that acknowledging the existence of passive euthanasia would have the bad consequences claimed, it is not clear that that would justify anything other than a piece of large scale public dishonesty. Consciously and dishonestly promoting the view that there is no such thing as passive euthanasia would presumably eliminate the bad consequences that this argument foresees. But it is hard to believe that the EAPC Task Force would support this policy. It is (at best) extreme paternalism, and the consequent failure to respect patient autonomy (as well as that of everyone else interested in the debate) would surely make it very unattractive to those who subscribe to the values standardly associated with palliative care (and health care more generally). And in any case, this is not the original Task Force view, which was not that we should pretend that passive euthanasia does not exist but rather that the expression passive euthanasia genuinely is a contradiction-in-terms.

Finally, a third objection is that we may have concerns about the possible bad consequences of denying the existence of passive euthanasia in this way. For once we start down the road of trying to secure our desired outcomes by a process of “conceptual cleansing”, the scope for political abuse in this Big Brother approach to language is obvious, as is the general concern for freedom of speech which such linguistic fiats raise. It would be better, we feel, to refrain from trying to shape people’s beliefs by manipulating the concepts and terms available to them, and instead seek to persuade them by substantive rational argument.

A DIAGNOSIS

Each of the arguments discussed above is open to decisive objection, so the EAPC Ethics Task Force view seems unsupported. But rather than (or at least prior to) dismissing it, it is worth trying to see why this highly implausible view might have been embraced in the first place. What attractions might it have had for people who are closely involved in the care of the dying, and hence have to deal with the (sometimes quite terrible) situations in which the practical question of euthanasia arises?

One likely cause of professional hostility towards passive euthanasia is an overly broad understanding of it. As we saw earlier in this paper, at least three conditions must be met for passive euthanasia to occur, and a failure to realise the significance of this, especially the importance of intention, may easily lead to the erroneous inclusion of very different reasons for withdrawing and withholding life-prolonging treatment (such as cost-effectiveness or futility) under the general heading of passive euthanasia. If this mistake is combined with the common view that euthanasia is morally objectionable, plus the view that some cases of withdrawing or withholding treatment are morally permissible, then it may seem that the only way to preserve that permissibility is to deny the very existence of passive euthanasia. But withdrawing or withholding life-prolonging treatment can, as we have seen, be done for a variety of reasons, many of which have nothing to do with euthanasia, passive or otherwise. Once this is realised, and cases involving a direct intention to bring about death are properly distinguished from cases where this intention is not present, the idea of passive euthanasia need not seem anything like so threatening to healthcare professionals. In most of the cases which they want to class as morally acceptable they may safely do so, without any risk that they will find themselves also inadvertently endorsing something they may strongly object to, since these cases do not have the intentional structure which is essential to euthanasia.

A few cases may remain which we are inclined to believe are both morally permissible and genuine cases of passive euthanasia. For those who are committed to the belief that all forms of euthanasia are wrong, there is nothing for it but to engage in substantive argument to determine which of these beliefs has to give way. But this is what ethical reflection is like: if we want to make our moral judgements consistent, non-arbitrary, and rationally justified, we have to engage directly with the arguments which may support or undermine them. We cannot settle these matters by conceptual stipulation, since what our concepts are, and whether there is anything which answers to them, are not things which can be settled by fiat. Claims about the nature of our concepts, and about what actions fall under them, are as much in need of supporting arguments as claims about the morality of those actions, and attempts to bypass the arguments leave all the important work still to be done. There is no escaping the need to examine the cases where we seem to agree on the application of the concept, the cases in which we are inclined to disagree, the nature of closely related concepts, and the difficult counterexamples to our proposed accounts or definitions. We cannot dispel our (understandable) uneasiness about passive euthanasia by announcing its logical incoherence, and arguing through stipulative redefinition is an attempt to achieve for free something which can only be acquired through genuine ethical endeavour.

Thus we conclude that the EAPC Ethics Task Force’s hostility to the very idea of passive euthanasia, especially their claim that it is “a contradiction in terms”, is unwarranted.⁵ Although healthcare professionals’ nervousness about the concept of passive euthanasia is understandable, there is really no reason to abandon it provided that it is properly and narrowly defined and provided that “euthanasia reasons” for withdrawing or withholding life-prolonging treatment are carefully distinguished from other reasons. Indeed, we would argue that passive euthanasia, when correctly defined, is a useful expression and one which healthcare professionals should be allowed to employ in their discussions of policy and clinical practice.