In the article, they offer three scenarios. You are to decide whether the protagonist's proposed action is "obligatory," "forbidden," or "permissible."
1. A runaway trolley is about to run over five people walking on the tracks. A railroad worker is standing next to a switch that can turn the trolley onto a side track, killing one person, but allowing the five to survive. Flipping the switch is ______.
2. You pass by a small child drowning in a shallow pond, and you are the only one around. If you pick up the child, she will survive and your pants will be ruined. Picking up the child is _______.
3. Five people have just been rushed into a hospital in critical care, each requiring an organ to survive. There is not enough time to request organs from outside the hospital. There is, however, a healthy person in the hospital's waiting room. If the surgeon takes this person's organs, he will die, but the five people in critical condition will survive. Taking the healthy person's organs is _______.
Hauser and Singer tell us that more than 90% of respondents agree on the answers to each of these questions -- actually 97% agree on the answers to 2 and 3 -- regardless of whether they are religious, or what religion they adhere to if any. (By the way my own answers were with the consensus in all cases.) They give this as a refutation of the assertion that many people make, that morality flows from religion and that people without religion must somehow be amoral. Instead, they argue, ethical intuition is a product of evolution.
I'll buy that, but what interests me about this test is that the preferred answers to 1 and 3 are different -- but if you look at them a little harder, it's very difficult to say what is different about the two scenarios. Neither protagonist has created the situation in which five people are imperiled. Both protagonists have the opportunity to perform an affirmative act which will save five and kill one. Both are acting in their professional capacities. Both potential martyrs are innocent bystanders. The difference seems to lie, somehow, in the degree of agency involved -- or perhaps some of you have a different explanation.
In ethical dilemmas, it's surprising how often very small differences in circumstances tip big differences in our judgments. Doctor Kevorkian helped dozens of people to take their own lives, and juries repeatedly refused to convict him -- until, in the final case, he actually physically opened the IV himself, because his client (or patient or customer or whatever he was) was paralyzed and could not do it. The client clearly stated on videotape that he wished to die, that he would open the IV himself if he could, and that he wished for Dr. Kevorkian to do it for him. Even so, that tiny distinction -- providing additional, wished for assistance -- was too much for the jury, which convicted him of murder.
The distinctions among allowing someone to die by turning off life support; providing analgesia with the stated intention of relieving pain, but knowing the dose may hasten death; and giving an injection for the purpose of hastening death seem very small and in a credible analysis, meaningless. But they are pretty much universally accepted. Action one is generally considered permissible under certain circumstances; people are somewhat divided about whether action two is permissible but it is de facto legal in the U.S.; and action 3 is illegal almost everywhere although some people feel it should be permissible under some circumstances. Note that unlike the scenarios given by Hauser and Singer, these situations do not produce an ethical consensus.