by Jeroen Bouterse
We are fortunate to live in an equitable society, with high levels of trust and excellent administrative transparency. Many old diseases have been eradicated, and most people live healthily to old age. However, our utopia has been suffering for a long time from a peculiar and cruel illness, about which after a century of research the following is now known with moral certainty.
Everyone has a one in five chance of falling victim to this disease. No genetic or lifestyle factors predict it in any way, and it is undetectable until you are exactly thirty years of age. At that point, there are two days during which it is symptomless but detectable with advanced medical equipment, after which it is irreversibly fatal and kills rapidly. There is no cure except one: during these crucial two days, tissue from the vital organs of one healthy peer can save with certainty the lives of five diseased people. However, the required procedure inevitably kills the donor.
A few decades ago, a voluntary association was formed. If you sign up, you commit to reporting to a hospital when you turn thirty, together with thousands of peers who have turned thirty on the same day. You get tested. If you are healthy, there is now a 5% chance that you will be randomly selected as a donor, which means you will die. If you have the disease, you will be saved. For both groups, the procedure is painless. You can decide whether you want to be informed of your predicament at any point, but you cannot opt out after entering the hospital.
This ‘5-to-1’-scheme reduces every participant’s chance of dying at thirty from 20% to 4%. This being an enlightened age, by now almost everyone has realized this is the sensible thing to do. The healthy participants who die as donors are not remembered as murder victims, nor praised as heroes, but mourned as casualties of the disease itself – their deaths regarded as equally indiscriminate and senseless, but fortunately fewer in number than they were in the past.
Systems verifying that everyone has signed their papers have been exceedingly strict from the start, as have the procedures ensuring that donors are randomly selected. In several decades, no mistakes have been made. But you never know! In a perfect storm, doctors might need to act under a bizarre combination of knowns and unknowns. For instance: because of a failure of some computer systems but not others, they might be certain that they are confronted with five diseased people who need to be saved quickly, and one healthy one. At the same time, they might not have timely access to the records confirming that all patients are of sound mind and have consented to the procedure, and that this is what they came to the hospital for, rather than a sleep study.
Given the robustness of our systems, such scenarios are sometimes waved away as outlandish – especially given how few people are outside of the 5-to-1 scheme to begin with. We are reminded of cultural fears in the distant past, about the split-second decisions self-driving cars supposedly needed to make. These seemed relevant at a time when they shared the roads with humans, who were still driving and regularly crashing their vehicles into each other. Even under that curious interplay of responsible agents and erratic forces, high-stakes ethical dilemmas under precisely the right amount of knowledge and ignorance were extremely rare in practice.
However, most people understand that behind such fanciful scenarios there is a question of principle. If you know with certainty that you can save five by killing one, but you don’t know with certainty that the one person has consented or would consent to be used that way, then can you make the decision for them?
There is little controversy around the answer. All ethics boards vote in favor of killing the healthy patient if there is no further information about everybody’s preferences, provided their presence in this situation is believed to be sufficiently arbitrary. (I.e., we cannot have reason to suspect that something about them – their curiosity, their clumsiness, or their connections – made it more likely that they rather than someone else were in this position.)
This consensus does not necessarily reveal its own operative principles, because it is overdetermined: the 5-to-1 scheme clearly and uncontroversially produces the best possible outcomes; under the ‘veil of ignorance’ a rational person would always elect to be part of a 5-to-1 society; and as regarding this matter we actually live with an almost ideal veil of ignorance, we know that random thirty-year olds likely have or would have consented even if we can’t get individual confirmation. Given all this, most people believe that canceling the procedure – allowing four more deaths – would be immoral and rightly lead to outrage.
Some individual ethicists have dissented, grumbling that rather than doing critical moral thinking, the boards are choosing the route of least resistance and sanctioning whatever course of action most resembles broadly accepted practices. These dissenters warn against any relaxation of the rule that requires a signed consent form: else, they say, there is a slippery slope towards killing people who we know have not consented, or who cannot consent.
At this point, hardline consequentialists have injected a thought-experiment into the debate, to illustrate that consent is not necessary to begin with. Imagine a society just like ours, they say, plagued by the same puzzling disease as ours, equally random and merciless, with the same single remedy in which one healthy person takes the place of five diseased ones of the same age – except now the illness manifests not at thirty years old, but at three.
Even in our halfway-utilitarian society, some feel uncomfortable performing cold, rational calculations when it comes to the lives of children. However, most people say that on their behalf they would make the same choice they have in fact made for themselves in their own lives. How could they bear to live in a world where 20% of young lives were cut short rather than 4%, just because those involved could not yet understand what the correct decision was?
Historians add that even in those dark ages when people were squeamish about self-driving cars, parents had their children vaccinated despite small but nonzero risks. In fact, they were often required to do so. Apparently, even then many people realized that low odds of dying young are better than high odds of dying young. Moreover, states ruled that not only the numerically literate should benefit from those better odds.
Then as now, changing the odds from worse to better often meant re-rolling the dice, in the sense that the people who died in the new system were not a subset of the people who would have died in the prior circumstances. We understand that this does not change the moral calculus. Of course, nobody wants to be that unlucky person that dies from the re-roll; but if the outcome of the first roll is unknown, almost everybody wants to get in on it – wants to live, that is, in a society that minimizes the number of lives cut short in a fair way, and maximizes their own chances of survival.
In the past, not everyone was able to shake off confusions around probabilistic reasoning. Idealized thought-experiments did not always clear it up either, especially if they stayed in a medical orbit. Somehow, people remained biased towards the notion that in the ‘default’ case, they wouldn’t die. Feeling vigorous now, they figured they would be that single healthy person in the hospital, not one of the five hanging by a thread in the operation room.
Interestingly, they did understand the principle in a different context: that if six people were distributed unevenly over two tram tracks, you would be more likely to be in the larger than in the smaller group. Indeed, an overwhelming majority figured it would be morally sound, perhaps even obligatory, for an untrained bystander to pull a lever that killed a single person over five. No consent forms involved! No checking computer records to see that mister Other-Track had opted into this classic 5-to-1 scheme. No, just a basic moral principle: fewer premature deaths are preferable over more premature deaths.
Their unwillingness to apply that framework to our situation was due not simply to their general state of moral confusion (evident from their xenophobia, short-termism about climate and resources, lackluster response to crimes committed by their allies, and treatment of animals), but also to the less clear-cut medical situation in their time. Most of all, it flowed from the far lower levels of social trust and organization they had to operate under. Under those circumstances, it was almost impossible to kill a random person to save five in any reasonably certain and procedurally fair way. In fact, it is remarkable they even came up with the idea. Had they tried, distrust would have skyrocketed.
Understandably, it was difficult for them to disentangle the ethical core of the situation from the contingencies of their own social and political context. To them, it just looked like a terrible idea, which it was. They rationalized this intuition in terms they thought were ethical – distinctions between ‘killing’ and ‘letting die’, and supposed rules around means and their relations to ends – but which were either irrelevant or wrongheaded.
In their clear-eyed dealings with the more abstract trolley situation, however, we can see that even they would have been able to understand, with the proper context, that the tendencies in our own society are morally sound. The 5-to-1 scheme has amply proven to be extremely beneficial, and the presumption has rightly become that all else being equal, ‘the’ healthy patient – meaning a fairly chosen arbitrary person – is to be killed. That we don’t make participation obligatory is because this would make an ever-smaller minority extremely unhappy, because their non-participation only harms themselves, and because our high trust levels are partly maintained by liberal laws with a strong aversion to compulsion.
Still, a nudge in the right direction remains warranted. At least, let’s reform the 5-to-1 to include default enrollment.