Helen Frowe wrote me yesterday to try to understand better my position on how to count the agent’s interest in a trolley switching case. The text she was trying to understand was a piece I co-wrote with David Wasserman, called “Agents, Impartiality, and the Priority of Claims Over Duties; Diagnosing Why Thomson Still Gets the Trolley Problem Wrong by Appeal to the ‘Mechanics of Claims,’” Journal of Moral Philosophy 9 (2012). Thinking about how to answer her brought me to consider an interesting case I hadn’t really thought about before. So I post it here on Pea Soup to invite replies to my tentative read on how to handle this case.
First, some background. I assume that a bystander at a switch is permitted to turn a trolley from five onto one, but is not required to do so. The question is why. Here’s what I wrote (pp. 554):
Negative agent-claims have weight, because they serve to protect an agent from having to treat herself as a mere tool for the promotion of the general welfare. This is not to say that they always tip the balance in their favor. If the sacrifice is small, and you would not have to make it so often that it would, in the aggregate, become substantial, then your negative-agent claim may not tip the balance. For example, if you had to turn a trolley away from five [to save them, and doing so would turn it] onto an empty track, and [if] you could do so without any great cost to you, then your negative agent-claim to be free not to do so would lose in the balance. You would be required to turn the trolley. But if the situation is as originally described in Trolley Switch, that is, if you would kill one as a consequence of saving five, then you would not be required to turn the trolley. This is not because the negative patient-claim of the one outweighs the positive patient-claims of the five. Even if we assume … that the patient-claims of the five [claims to be saved] outweigh the patient-claims of the one [a claim not to be killed], you are still not required to turn the trolley. The explanation of this fact is that you have a substantial interest in not having to be an agent of death. That interest expresses itself in your having a substantial negative-agent claim not to be required to turn the trolley onto the one. You might waive that claim and choose to do so anyway, but you can also rely on that claim to ground a right not to do so.
Helen wrote me to ask whether the reason the agent claim is strong enough to give you a right not to turn the trolley reflected the objective value of not being an agent of death, or the subjective value you might (perhaps within reasonable limits) put on not being an agent of death.
Before I say how I answered that, I want to clarify a bit more about how I think the agent claim works. Helen wasn’t sure if it was meant to be a bonus for being free to act autonomously, to do what you want. The answer is no. It can’t be just that. Keep in mind that the same autonomy interest is at stake when the other track is empty and when it has one person on it. Of course, the fact that one person is on the other track also makes an important difference, and we shouldn’t suppose it’s simply by cancelling out one claim for aid; it might make a substantial qualitative difference. Nonetheless, I think the magnitude of your negative agent-claim must also matter, and that magnitude depends on the interest of yours protected by it.
To see that, suppose that you are the fat (or massive) man (or woman) on the bridge, who could jump in front of the trolley and stop it, save the five, but lose your own life. It’s permissible for you to do so; it would, indeed, be praiseworthy. But it’s a more or less unshakable fixed point for me, anchoring my search for reflective equilibrium, that you need not sacrifice yourself in that way, at least not you have done nothing to acquire a special responsibility for the welfare of the five. Not sacrificing yourself is not merely an excusable failure; it’s a right you have; doing it would be clearly supererogatory. Only the most hard-core utilitarian would deny that. The question is how to account for it. If you are just weighing up your interest in your life and your interest in your autonomy, as though they should just be added together, you would see that the interests of the five are greater. The only way you can give greater interest to your own life is if it is augmented by some sort of agent-centered multiplier.
But I want to be clear, I am not endorsing the sort of agent-centered prerogatives endorsed long ago by Samuel Scheffler and defended more recently by Jonathan Quong. It’s not the case that you can multiply your interests to do whatever you want. For example, you cannot justify turning the trolley from one onto five the way you can justify allowing it to kill five rather than turning it onto one. And the difference can’t be simply that claims not to be killed are that much stronger than claims to be saved. If claims not to be killed as a side effect of saving others were that much stronger than claims to be saved, then it would not be permissible to turn the trolley from five onto one. Rather, the point is that you can only use the multiplier when your interests are negative.
The relevant principle is that you have a strong basic claim not to have to live as a means for the welfare of others—the exception being, again, if you have done something to acquire a special duty to them. And to be clear, negative claims do not really enjoy to a “multiplier,” as though it’s a linear function. The point is really that the space of rights has to allow us to lead our own lives. Small sacrifices, that would not be demanded so often as to become a big sacrifice when added together, can be required of us. But when even an individual sacrifice would be a significant cost to the agent, then there comes a point at which no matter how great the good, the agent has a right to refuse. I am sure that this is true for having to give you life to save others to whom you owe no special duty. But being an agent of death, that seems more complicated.
With that, I return to Helen’s question about whether the strength of the negative agent-claim depends on the objective magnitude of the interest protected, or the subjective value you might (perhaps within reasonable limits) put on it. To test intuitions, consider this case: Suppose I do not in fact place a lot of value on not being an agent of death, but I haven’t waived the claim. Imagine the case of an army sharpshooter. I’ve killed, say, hundreds of people when serving as a soldier. I didn’t think of them as necessarily bad people who in any way deserved to die. There were simply fighting for the enemy, and I therefore took them to be legitimate targets (I haven’t read or been convinced by McMahan about their actual liability). I’ve also dropped bombs, killing innocent bystanders in the attempt to destroy a proportionally important target. And I’m not done. I’m about to go out to serve another tour of duty next week. So I’m a conscientious guy, I try to act morally, but I don’t mind being an agent of death for a greater cause. And then, on a weekend at home, off duty, I find myself being the bystander at the switch. Can I claim the right not to turn the trolley?
My intuition (which, with about $4.50 will buy you a latte at Starbucks) is that I have no duty to turn the switch. Contrast a guy who works for the trolley company, or a police officer. I think both have a duty to turn the trolley. They left their personal interest at home when they came to work that day, at least when it comes to choices like that, choices that fall in the line of duty, as it were.
My defense of this intuition, insofar as I want to defend it, is that prior commitments and actions can change what is permitted, but only by giving agents special responsibilities. They cannot change what is permitted by showing agents to have lost the option to express a value that others can reasonably embrace. In other words, we should embrace an objective value account of the interest at stake to set the strength of the agent-claim.
On the other hand, consider this. At some point as the numbers grow bigger in the save column, it starts to seem more and more dubious that the agent at the switch can refuse to throw it. Assume that 50 to 1 is that tippy point. Now it is hard to imagine the person who has killed so many and doesn’t really care about killing others invoking the right not to have to become an agent of death one more time. Now that seems like lazy, selfish, indulgence. Maybe he has to turn the trolley. But one should be more sympathetic with someone who has been a committed pacifist for years. I think he can hold out for bigger numbers in the save column. But that indicates that it is the subjective value that the agent (reasonably) assigns to not doing something that determines the strength of a negative agent-claim.
It could be that a hybrid is correct: that we should use a baseline objective measure of the agent-interest at stake, and that subjective valuation can add to that, creating a stronger claim that will hold out against stronger competing claims to act.