Statistics from Altmetric.com
Recently, Bonnie Steinbock has argued that there is still not a convincing case to support the legalisation of doctor-assisted suicide.1 The argument is framed in consequentialist terms: rather than contend that there is something intrinsically wrong with mercy killing itself, caution is recommended because of the risk that a system may be open to sufficient abuse to warrant its non-implementation. A welcome criticism is made of partisanship that obstructs useful progress in the debate, which she suggests should be based on objective, empirical evidence on both the need for, and the risks associated with, legalised doctor-assisted suicide.1 The main presumption underlying Professor Steinbock’s argument is that a proportionately large number of people would become vulnerable to great and unacceptable harm under a regime that was not strictly regulated. She asserts certain values, which, on a consequentialist calculation, support her conclusion that currently the case for legalisation is not made.
Here I would like to question how useful objective, empirical data can be. Although data that show some assertions to be objectively right or wrong may be obtained and are doubtless valuable, I will show that the debate on assisted suicide is unavoidably partisan and that the protagonists will therefore receive data differently, necessarily because of their moral and philosophical convictions. Steinbock would like the question of legalisation to “turn on objective, empirical evidence about both the need for PAS [physician-assisted suicide] and the risk of mistake and abuse.”1 But to be understood, the concepts of “need” and “abuse” require pre-existing value judgements. Furthermore, even if the values of these concepts are quantifiable, they are probably parts of distinct metrics, so offsetting one against the other will become a matter of qualitative assessment. This does not entirely disvalue the use of objective data, but it means that the emphasis of the arguments for or against assisted suicide must be rooted in ethical argument, with ethical reasoning for choosing one legal path over another. Despite Steinbock’s appeal to rationality, I shall argue that there is insufficient ethical concord to allow decisions to turn solely on questions of empirical fact.
PROBLEMS WITH PRESUMPTIONS
Expediency demands many things to be taken for granted in most arguments. It is often possible to ascertain only the most important points: normally through assertion followed by explanation and, ideally, justification. Why? Because it is possible to say “why” to almost everything, and the result of unlimited responses to it would be unlimited argument. Usually, a debate benefits from its focus and limited length because much is already established as a starting point, and if that is wrong it can be dealt with separately. The interconnectedness of things, of course, makes practice less simple than theory as just stated, but the point stands.
So, this is not a criticism of taking some things for granted for the sake of argument: it is accepted that as much is necessary. Rather, it is a caution arising from that fact. By looking at a couple of points presupposed in Steinbock’s paper, we can observe an attachment to some moral norms that are neither objectively nor empirically proved. The relevance of this is that they can only contentiously be taken for granted and will require strong justification. This is not an attack on Steinbock’s ideals, but a suggestion that her paper serves as evidence that non-partisanship, at least in some areas, is unattainable. Her paper was not the place to defend at length her entire ethical stand point. But while I would embrace escape from unwavering, rhetorically loaded bias, I do not think that the question of assisted suicide’s acceptability can be limited to objective standards or to the data inferred from empirical observation. The facade of the self-evidence of certain principles or assertions should be acknowledged as an inevitable aspect of this debate. Otherwise, the gloss provided by attachment to claims of seemingly pure objectivity will either overstate or discredit an argument that would be strengthened by acknowledgement of its inescapable imperfections.
The first point here is about understanding suffering. Suffering in this context is generally assumed to have an association with physical pain. It is not the case, however, that human suffering is limited to, or even necessarily caused by, any physical stimulus. It can be the product of, or exacerbated by, people’s evaluation of their circumstances. Steinbock implies that the emergence of a duty to die would be bad. Why? Because the reasons for the duty may be bad, it seems, given the rhetorical questions that are posed.1 The idea of a legal duty to die seems utterly repugnant. But a moral duty to die, or the feeling in some members of secular society that such exists, is not beyond reason. It must at least be argued against, rather than discarded out of hand: if the prolongation of life is not to be afforded absolute value, its very high value must be open to subordination. An unexplained blanket refusal of its potential to be less important than “serious financial, physical or emotional burdens on family members”1 is troublesome and it leaves hugely restricted the question of what is worse than death for a social being. As Peter Railton2 states: “When one studies relationships of deep commitment—of parent to child, or wife to husband—at close range, it becomes artificial to impose a dichotomy between what is done for the self and what is done for the other.”
I would argue against a deontological approach to ethical issues related to assisted suicide. But discussion on legislation is inevitably tied up with talk of rights and duties. So, although I would join Steinbock in rejecting arguments that it might become a person’s (legal or moral) duty to stop living, the case must be made. To take its conclusion for granted is to treat it wrongly as objective reality. This is especially so because even if a duty to die is rejected, a rejection of the applicability and acceptability of the contribution of external elements to a person’s suffering asks for an objective account of experiencing suffering that either does not exist or cannot be determined. It may be answered that an accurate determination can never be made, whatever is taken into account. I would accept that, but maintain that we get closer to objective truth by not denying ourselves consideration of what relevant factors may be.
The second point I would like briefly to look at here is the presumption that a physician must be involved, partly, to provide some means of policing family members with bad intent. Steinbock alludes to the “obvious reasons” for no jurisdiction explicitly permitting relative-assisted suicide.1 Omitting to state obvious reasons makes sense, but presupposing plainness does not. It is possible that an assisted suicide system in which the role of the physician was not pivotal could work well. Serious consideration is needed, supported by empirical research where it is possible, not a closed door to the idea. Steinbock might be interested to look at the situation in Switzerland, where (non-doctor) assisted suicide has been lawful for more than 60 years. Article 115 of the Swiss Penal Code permits assistance with suicide by anyone who acts with entirely honourable motives. The presumption that a system in which relatives are allowed to assist suicide would obviously be bad rests on an assumption—possibly a correct one—that human nature is overwhelmingly unworthy of trust, or that non-medics cannot properly assess a request for help in death. It is not necessarily so, however, and the Swiss law’s longevity evokes prima facie scepticism of the cynicism. It would be more in line with the professed message of Steinbock’s paper to seek evidence from Switzerland, and not to rely solely on intuitive, rather than objective, reasoning. (Of course, as Steinbock1 notes, even if a system works in one jurisdiction, it does not automatically follow that it will work in another.)
The acceptability of arguments regarding assisted suicide is most often a question of ethics. Arguments from, for example, economic views are generally rejected not because they are wrong per se, but because an economic analysis, at least on its own, is not thought to be the right framework within which to hold the debate. Steinbock considers some of the ethical arguments for and against doctor-assisted suicide, before stating that the consequentialist account that she endorses is concerned with the social consequences of legislation. However, the goods and bads that are involved in her reasoning seem to be ethical. A substantial part of the basis of her current rejection, indeed, is the fear that society may further its own ends by abusing vulnerable groups in response to economic pressures.1
Religious arguments against doctor-assisted suicide are rejected on the grounds that their use is not justified in a secular society.1 No explicit answer is given, however, to modified arguments based on, for example, Christian ethics. There are many commentators who contend that society would be ill served by any regime that permits active help in killing, and who believe that this is true even in secular society.3–6 If the reasoning is seen as a question of ethics, it becomes apparent that the criteria which Steinbock suggests justify mercy killing need to be defended. They become contentious conclusions of ethical debate rather than useful starting points in a calculation. A particular reader’s sympathies may lead him to agree with Steinbock’s opinions and analysis, but what of a reader sympathetic with the values of, for example, Margaret Somerville5 or John Harris7? Depending on his value system, reader’s interpretation of any empirical data will be subject to variation, even if he accepts them to be facts.
A further question that arises in relation to Steinbock’s ethical analysis is why a statute would have to cover “all and only the justifiable cases” (emphasis added).1 A rational consequentialist would surely feel compelled to legislate if the net outcome would be at all better.8 That is to say, if some people would benefit and everything else were equal, the fact that some justifiable cases would still not be helped would not justify inaction. More contentiously, if the calculation meant that an amount of harm would be done because some unjustifiable deaths would occur, but overall more good were achieved because sufficient justifiable killings were achieved, then legislation should be advocated. This may seem loathsome, but it is a position that a non-absolutist must accept or, at least, reject with reason, if the outcome of his consequentialist calculations demands or seems to demand it.
As Steinbock acknowledges, the relevance of data collected in a different society may not be easily determined.1 This means that the arguments are bound to be speculative, and therefore lack a completely convincing objective basis. If we are able to offer empirical support to an assertion by reference to relevantly similar cases, there is some value, of course. But the success of another state’s regime will be determined partly by the values of those who devise and police and use it, and its relevance for a third party jurisdiction will not be objectively ascertainable.
A problem exists with the notion of evidencing objectively the need for assisted suicide and the terms that would mark its misuse. If a position is framed, we can use it as a scale against which objective data may be measured, but the position itself will be a human construct.9,10 A look at the breadth of disagreement within the debate on mercy killing shows how a widely acceptable measure will be nearly impossible to create. It is questionable whether those who disagree with Steinbock’s ethical premises will attach much value to evidence that cogently shows that these premises are unlikely to be flouted by a certain framing of a statute. Perhaps we do not seek to convince everyone, and Steinbock does not necessarily claim objectivity in her own ethics themselves, but for any useful form of agreement to be met, I would suggest, we need to investigate more fully our ethical positions. The main value, then, of empirical data will be more to help prove or disprove matters of fact—for example, to add weight to arguments for or against the inevitability of an empirical “slippery slope” if assisted dying legislation is passed. Even then, though, a claim of the data’s objective strength will not be absolute, as speculative reasoning will also be involved in the application of the findings.
The final practical problem that I wish to mention is the difficulty with metrics. Even if everyone were to accept the values that Steinbock advances, the measure of need—seemingly ascertained by reference to a terminally ill person’s understanding of his position—and the measure of abuse—seemingly ascertained by reference to the disvalue to human life and the consequent malign effects on society—are probably not marked on the same rule. A coherent equation involving both may just be out of the question. And quantification of each, even if they are measured on the same scale, would surely not be a simple, objective matter. The assessments themselves, at least in the measure of need, require in large part the subjective judgement of each person.
I should point out that a consequentialist approach is one I would endorse, but I do so with the caveat that the assessment would be one that is in large part based on subjective values, and the balancing is achieved by qualitative assessment of harm and benefit. An argument that posits itself as being based in objectivity sounds very convincing: if it is in fact only as objective as possible, there can be a marked difference. If cautious understanding of legislation on assisted suicide is well advised, so must be cautious understanding of that understanding. A protagonist must be prepared to defend his values, and question the validity of what to him may seem obvious.
I have not sought to discredit the beliefs on which Professor Steinbock’s arguments are based, and I accept that it would not be possible to give an explanation of each premise in her argument. I have shown that partisanship—even if inadvertent—is unavoidable. Even having “no position” is in fact a position: it is a reflection of values that include an acceptance that the values may change or may not be known. I agree with the argument that interpretation of fact is susceptible to spurious understanding dictated or directed by a party’s entrenched position, and I would welcome a minimisation of the practice.
What Steinbock successfully shows in her article is that if her values are accepted, the best means for testing the acceptability of proposed legislation is with consequentialist calculation, involving empirically proved figures. What I hope to have shown is, firstly, that the calculation necessarily involves contentious and objectively uncertain value judgements, thus reducing its strength, though not fatally. Secondly, a degree of speculation will still be involved in any decision making, even when objective data are used. I also hope to have shown that some suppositions should not too readily be taken for granted. Much remains to be resolved in the ethical debate. There is value in using data to add weight to arguments for or against, for example, the existence of an empirical slippery slope, but if a consequentialist account is to provide acceptable justification for legislative decisions, we first need agreement on what value should be attached to the elements of the equation. Until the meanings of need and abuse are established, a consequence-based conclusion will compel only those who already agree with it.
I thank Professor Søren Holm for his comments on an earlier draft of this paper and for his valuable points made in discussion.
Competing interests: None declared.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.