Puzzles: The Problem of Harmfully Using People

35 Replies to “Puzzles: The Problem of Harmfully Using People”

1. Thanks for the puzzles Victor! Here’s my start at some answers:

   Jump Behind is permissible. Even if one adopts a causal view of the means principle, one can think that there is a difference between creating the causal structure that allows one to use another as a means and exploiting a pre-existing structure. (I call the latter a “beneficial opportunity” in my review of Quong’s book.) We see this distinction in other cases, including Ketan Ramakrishnan’s Room case (where the one person’s body enables access to the five) and Alec Walen’s On the Button case where the one person on the sidetrack is stuck on a button that enables the trolley to be switched onto the one. In all cases, the actor just exploits the pre-existing causal structure. Still, in Jump Behind, Bella is permitted to move herself out of the way if she can (as can the people whose presence is exploited in the other hypos). Pull In Front is impermissible because one treats another person as a tool in one’s toolkit to be deployed for one’s use (or the greater good).

   I find Jumpers more perplexing because it is a competition case with the means principle added in. I am not convinced I’d resolve it any differently than other competition cases (like running towards alcoves)… but I can be convinced.

2. All of these suffer from the problem that I don’t know whether or when hitting one person would be enough to stop a dangerous trolley. Given the kinds of things trolleys are, I suspect that nobody knows it. I’m not even sure how to imagine that I did know. There would necessarily be lots of uncertainty involved. So the stipulation of the thought experiment asks me imagine a situation which I can’t clearly imagine, so much unlike the actual world that I can’t trust any pumped intuitions.

3. Thank you, Victor, for these interesting thoughts and time and effort setting out this puzzle. It’s an insight into the analytical philosoper’s method and mind, now found on the web. I’m less convinced it is found in or has as much relation elsewhere in the world.

   These scenarios, like many a ‘trolley problem’, mix things we probably all know well with other things we possibly never could, as PD above identifies. To set that out further, for example, that bears and trains can both kill you, along with in the first scenario that Bella can’t jump, in the last that Yogi stops (or slows sufficiently) after one good mauling or meal. A lot of trolley problems seem to me to ‘gain traction’ from imputing everything that is posited to the minds of the participants when they are things they could not possibly know.

   Where you suggest the view in jump behind that ‘the only reason for [Archie] to jump is that Bella will then be used for his survival’, I am not so clear that this actually is a reason for him. Don’t we just have people trying to get as far as they can from oncoming (bears and) trains?

   The greater good does not seem so clearly to be involved here but we could possibly modify the scenarios so it is. We could make Bella pregnant, for example, so that she is great with Child, and ask whether B is permitted to pull A on the tracks to save herself and C. But in reality we still come to something like ‘the safer way’ doctrine as found in (the good sort of(?) casuistry, i.e. when we don’t know we are better not to act.

   We could also specify more, like have B in Jumpers tied to the tracks, have a small and readily stopped trolley (but with sharp edges) or a moderately lazy, only moderately hungry bear. But then these become both less ‘intuition-pumping’ and less obviously applicable to the world.

   It’s entertaining to think abut the contours of this but I wonder if I could ask whether you can think of somewhere this puzzle might actually present itself?

4. Hi Kim,

   thanks very much for your response to this. One difference between Jump Behind and Pull in Front is that the former is intuitively an allowing case, whereas the former is intuitively a doing case – that, at least, are my metaphysical intuitions. Even if Archie acts in a way that results in Bella being harmed, he merely allows her to be harmed rather than harming her. Things are otherwise in Pull in Front. Your explanation didn’t refer to doing and allowing, but I wonder if it is really just a version of that? (that was my solution to these kinds of case in To Do, To Die, To Reason Why).

   Of course, a familiar view about versions of the Means Principle, is that it applies to allowing as well as doing – many think it wrong to allow a person’s illness to develop to see how to cure it where several other people will be cured as a result (Quinn’s Guinea Pig Case). But, you might reply, even if that is wrong, it is easier to justify allowing the symptoms to develop than giving the person the illness. I feel the force of that. So, did you mean something other than Doing and Allowing, and do you think that this also applies in the Guinea Pig Case?

   If you agree that it is wrong to allow the patient to die in the Guinea Pig Case, though, but that it is permissible for Archie to jump in Jump Behind, it seems important that Archie will be useful if he doesn’t jump; that would distinguish the case from the Guinea Pig Case, where the 5 will not be useful if they are not saved.

   PD and James, you rightly point out that my cases are highly stylised, assume full knowledge, and have no important implications for any real world problems. You see this as bug; I see it as a feature! They help us to get clear about the significance of using without being distracted by the pesky real world!

5. Thanks Victor.

   Here’s some half-baked musings:

   I’m not convinced that Archie can jump behind in Jump Behind. In doing so he avoids being used, but then uses Bella in exactly the same way. In the same way that (all else equal) we cannot do x harm in order to avoid x harm, I wonder if we should say that (all else equal) we cannot use another to prevent ourselves being used in the same way and to the same degree.

   Kim says that in jumping behind Bella Archie ‘exploits an existing causal structure’ but I want to know more about how to understand this. Consider this case:

   Out of control: Bob’s self-driving car has malfunctioned and is hellbent on crashing into Carol. If Carol jumps behind Daphne at the last minute, then the car will hit Daphne.

   Is Carol permitted to jump here? If not, how does this differ from Jump Behind? Daphne isn’t on a track, like Bella, but how morally relevant are train tracks? In both cases Bella and Daphne are initially safe. Jumping puts them in danger. The causal structure as was was going to see them unharmed. The only difference is that Carol wasn’t initially protecting Daphne. But as I said earlier, while I think Archie is not required to allow himself to be used, he cannot avoid being used if the only way to do so is to use another in the same way and to the same degree.

   Here is one way to get at what I’m trying to say. When it comes to allowing, we generally have greater latitude/personal prererogative. I can justify allowing you to suffer [more than] harm x in order to avoid harm x for myself. When it comes to doing, we have less latitude. I cannot inflict harm x to avoid harm x for myself. In fact, I have to be avoiding substantially greater harm to myself. (Assuming innocence etc). The question then is whether the using principle behaves like doing or allowing. I am tempted by the view that it behaves like the doing principle—it’s a stringent principle, and so one cannot avoid being used if this means using another in that way.

   What then should we say about Jumpers? Does the fact that there is mutual using, or attempting to use, make it ok? I am tempted to say no. Archie uses or tries to use first. That wrongs Bella. (So maybe Bella is justified in jumping, to reset things, but Archie isn’t).*

   I am less sure, then, about what to say about Bear. My intuition is you can run. But i wonder if the *way in which* the other is used or exploited here is different enough that we can explain an important difference here. In Bear, the faster runner exploits the presence of the slower. in Jumpers the person is used as a physical barrier, and so used as a tool or thing. This might matter.

   * I assume here they become aware of the threat at the same time. Things get tricky if they become aware at different times. Not sure what to say then.

6. A chapter of my dissertation was on similar things and similar cases. If you’re interested, you can read it here: https://tjb207.user.srcf.net/me.pdf (the relevant chapter is “The Bystander Doesn’t Kill”).

   As for your puzzle about Jump Behind, I’d say he may jump behind and he may do so because, and unlike in Pull, he doesn’t thereby kill Bella. Just as he doesn’t in Jump. Just as the agents don’t kill in the cases I discuss in the chapter. (Or so claims I.)

7. Pingback: Navigating the Labyrinth of Morality: PEA Soup Launches Provocative New Series on Philosophical Puzzles – Kitab Dar

8. Thomas, thanks for that, and I look forward to having a chance to look at the dissertation. In the meantime, though, it seems that your solution presumes that allowing is all that matters, but Guinea Pig is an allowing case, and most think that the doctors ought to cure the one patient rather than allowing them to die so that they can cure the five.

   Thanks Patrick

   I suggested to Kim when discussing the Guinea Pig Case, if Archie is permitted to jump in Jump Behind, this must be because he will be useful if he doesn’t jump to distinguish it from the Guinea Pig Case. He would get to use Bella only because doing so would prevent himself from being useful. That would distinguish the case from Out of Control, where (I take it) he would not be useful if the car crashed into him, because he is not preventing the car from killing someone else.

   You suggest that Archie is not permitted to jump in Jumpers because Archie would be ‘using first’, but I wonder whether this is right. If he does nothing, he is usefully prevents Bella from being killed, and that would be his only reason for not jumping. If he was not useful, he should surely jump. So isn’t he useful first? And if so, isn’t he permitted to jump to prevent himself from being useful first? So doesn’t your principle suggest that Archie is permitted to jump behind, but if he does this, Bella is not permitted to then jump behind him?

   Another thing to think about with these cases. In the original Jump case, I think not only is it permissible for Archie to jump, it is permitted, or perhaps even required for neutral third parties to help him jump, even though doing so results in Bella’s death. If he needs help to get off the track, neutral people ought to give him a hand. I think this just as I think in the Guinea Pig case that neutral doctors ought to cure the one person rather than allowing them to die in a way that will result in the five being cured.

   For those who think that Archie is permitted to jump in Jump behind, is it also true that neutral third parties are permitted or required to help Archie? At least my initial intuition is that they are not. If that is right, it seems that if Archie is permitted to jump he is permitted to do so 1) to avoid being used; and 2) on an agent-relative basis.

   And one final question. In Guinea Pig, suppose that the 5 could either cure the one patient or allow them to die so that they could discover how they will be cured. Are the 5 permitted to allow the 1 to die, or must they act just like the doctors and cure the one. I am tempted by the view that they must cure the one, but I’m not sure. In Jump, if Bella can help Archie get out of the way, must she do that? Again, I’m not sure, but I’m tempted by the view that she must. However, in Jump Behind, I find it clear that Bella need not help Archie jump behind.

   Patrick, you could support your view that Archie is not permitted to jump in Jump Behind by arguing that agent-relativity is irrelevant in cases like these.

9. ‘He would get to use Bella only because doing so would prevent himself from being useful. That would distinguish the case from Out of Control, where (I take it) he would not be useful if the car crashed into him, because he is not preventing the car from killing someone else.’

   I get that’s a difference between Jump Behind and Out of Control in that it has the use or be used element… Out of Control was designed to push on the idea (introduced by Kim) of exploiting an ‘existing causal structure’. In what way is there an ‘existing causal structure’ in Jump Behind that is not present in Out of Control? That was my question.

   On the issue of use or be used (and on your question about who uses first in Jumpers), as I said, does the means/using principle act like the prohibition on doing harm or allowing harm? If it behaves like doing harm, avoiding being used doesn’t license using in the same way/to the same degree.

   The difference between Guinea Pig and Jump Behind is, I think, more complex than merely a difference between allowing to be used and allowing to be used to avoid be being used.

   The doing/allowing classification of this case is, in my view, complex. If Archie jumps in Jump Behind, he is allowing the harm (as Victor said earlier). But he’s arguably not – as in Guinea Pig – allowing Bella to be used. He is acting-so-as-to-use her. In other words, the doing/allowing classification when it comes to the harm comes apart from the doing/allowing classification when it comes to the using. Or something like that. At the least, the agential involvement in these two elements of the case seems different.

   And since jumping would actively use, this is why, to me, it follows the model of a doing harm case. We just allow harm x to ourselves over doing it to another. We must allow use Y over actively using another in that way. Otherwise I don’t think we’re taking the using principle seriously enough.

   And this better describes the difference between the two in Jumpers. Archie is useful to Bella at the outset of the case, but she is not actively using him (or allowing him to be used, as she has no control over whether Archie stays there, as the case is set up). But if he jumps, he is using her, and actively making it the case that he is using her.

   More generally, Victor, I wonder how all this fits with your well-known claim that we can only use others as a means when they have a duty to do so? If Archie is permitted to jump, he uses Bella. But surely she has no such duty to be used?

10. Thanks Patrick. On duties and means, generally think roughly this

    Means Principle: it is wrong for X to harmfully use Y, or to allow Y to be used, for the sake of g to degree h unless Y either consents or would be required to bear h for the sake of g.

    However, in cases like Jump and Jumpers, it may be that the following is true: 1) someone will be harmfully useful for saving someone else. And 2), that person, whoever they are, would not have a duty to bear the costs. In that case, the Means Principle does not give us the answer. That principle, as stated, will be violated either way, and that suggests that it is incomplete. We need a further qualification to deal with cases where someone will certainly be useful to save another (and also, I guess, cases where if no one is useful to save others, things will be worse for everyone). Exploring these cases helps us determine what that qualification is.

11. I wonder if we should distinguish between someone being ‘harmfully used’ and ‘being useful’.

    In Jump Behind, the default/do nothing option is that Archie is useful to Bella, but she does not ‘allow him to be used’ nor does she clearly ‘use him’, since she has no option to do otherwise. Arguably then she does not act wrongfully, and the principle is not clearly ‘violated either way’.

    Notice that your formulation of the principle concerns using or being allowed to be used, whilst your description of the case concerns someone inevitably ‘being useful’. This may matter.

    So perhaps the principle stands, Archie will violate it if he jumps. Bella will not violate it if Archie doesn’t jump.

12. Thanks, Victor. Just to push back a little but also as an attempt to contribute, without perhaps quite being engaged in the same enterprise… .
    You say “They help us to get clear about the significance of using”. Do you mean by this the significance of using in the real world, or the significance of using in a highly stylised world of full knowledge?

13. Hi Victor. Yes, my view is that killing (or not)* is what matters for the cases you raise in the post. Guinea Pig is a different sort of case, though not that different, since it concerns saving (or not). We ought not kill and we even ought not kill in order to prevent five other killings. Similarly, we ought not fail to save and we even ought not fail to save in order to prevent five other failings to save. I talk about this briefly in my “Saving and Letting Live.”

    *I loudly disavow understanding this or generalising it in terms of doing vs allowing. That is a bigger point, though one I discuss in my “Making Metaphysics.”

14. Hi Patrick,

    I’m not sure your suggestion in the second section could be right if it relies on the idea that Bella can’t do anything. Suppose we alter the case, so that Bella could pull Archie behind her. I take it that she is not required to do this. But her only reason for not doing this is that Archie being where he is is useful to protect her – if she were immune to trolley crushing, she ought to move him behind her, so her reason for leaving him where he is must be that she prevents the trolley from hitting him. But in that case, by not moving him, she allows him to be used. However, whether Archie is permitted to jump behind doesn’t seem to depend on whether Bella could move him behind her does it? If not, there seems to be no difference to the ‘being useful’ case and the ‘allowing to be used’ case.

    James – yes, I think we get a grip of the wrongness of using in the real world, full of risk and uncertainty, by getting a grip, first, on using in stylized cases where full knowledge is assumed. The reason is that risk or uncertainty of f almost always matters only because f matters. So to determine what we should do in cases where we are uncertain of f, we should first work, out whether, how and to what extent f matters. Given that risk and uncertainty are difficult topics quite independently of using, it seems unwise to try to consider more realistic cases that involve both using and risk and uncertainty together – it’s hard enough to develop a philosophy of using and a philosophy of risk and uncertainty independently of each other, and then put them together (considering potential interaction effects). Trying to do both things at once can’t be the way forward. If we don’t use stylised cases we’re just wondering around in the fog in the real world, with a myriad of potentially morally significant things going on that we can’t assess. One final thought – I just can’t understand why people think that uncertainty makes so much difference to our judgements anyway. It doesn’t in simple cases. Think about this dialogue:

    Me: It’s intuitively wrong for one person to shoot another in the face killing them for no reason.
    Objector: But in the real world, we don’t know that bullets will hit, or that they will kill people, so we can’t get reliable intuitions about your stylised case.
    Me: ???????

    Why is anything different in my more complicated cases? We don’t have such strong intuitions about these cases. That’s why they’re set as a puzzle. But I really doubt that’s because in the real world we’re not sure whether trolleys will kill people.

15. Thanks for the puzzle, Victor!

    It seems notable that a “don’t pull someone to their death” rule has millions of occasions for application every day, e.g. at train stations, bus stops and pedestrian crossings, whereas occasions for applying rules against jumping behind someone so they die rather than you, or running away from a bear that’s also chasing someone else, happen almost never.

    It might be replied that there are not many occasions for a more fine-grained “don’t pull someone to their death” rule, which would be something like “don’t pull someone to their death in order to save your life.” But the point is that there are many occasions for the coarse-grained “don’t pull someone to their death” rule, which encompasses all the cases covered by the fine-grained rule. Hence the coarse-grained rule is taken up through socialization, and it’s only that rule that’s needed to explain the intuitive response to Pull-type cases. There is no equivalent coarse, widely adopted rule for “jumping behind” or “running away from bears,” because you don’t have millions of commuters per day in a position to kill people by jumping behind them or running away from a bear.

    This doesn’t quite suggest an evolutionary debunking argument a la Singer (in evolutionary time, bears are more important than buses) but it is a kind of debunking argument – a social one. It suggests all the cases are the same in terms of fundamental morality and critical-level thinking (in Hare’s sense), but at the intuitive level, there are varying degrees of social (and maybe legal) acceptability, tracking the frequency with which certain categories of danger arise.

    To test this, suppose a country, Ursinia, was infested with bears, and a strong, universal norm against running away from bears in the same direction as other people was adopted. You can run away from bears, but only at such an angle that a bear following you is not thereby following another person too. “I don’t need to outrun the bear, only the person” is hammered home as the pinnacle of immorality. This socialization effectively saves many lives, as the bear doesn’t always follow the slowest runner and both runners therefore escape more often.

    Someone from Ursinia would hold that running in Bear is more wrongful than pulling in Pull, on account of their strong anti-running-away-from-bears-in-the-same-direction-as-people socialization. They stick with this judgment even when we clarify that there’s no option of running away at such an angle that the following bear is not thereby following another person, just as we’re inclined to stick with the judgment that the pull in Pull is worse than running in Bear even when it’s clarified that it’s not like randomly pulling a commuter because this time it would save as well as take a life.

    Could we honestly tell the Ursinian that they’re wrong, and that pulling just is inherently worse than running? On reflection I think we’d say that the contrasting intuitive reactions to the cases track nothing deeper than socialization. And if we’re moral realists, we could say this while thinking that other of our intuitions (e.g. random killing being wrong) DO track something deeper. Much the same would follow if we imagined a country where jumping behind people is the greatest danger.

    This kind of consequentialist response is perhaps not playing by the implicit rules of the puzzle. I hope this might be forgiven as the frequency aspect of the cases seems so prominent. I enjoyed the puzzle!

16. Have you read this? Boorse, C., & Sorensen, R. A. (1988). Ducking Harm. The Journal of Philosophy, 85(3), 115–134. https://doi.org/10.2307/2027067

17. @Michael Smith
    Thanks. I hadn’t read that, but this from the first page would certainly run afoul of Ursinian norms:

    ‘In a recent joke two campers, Alex and Bruce, meet a ravenous bear.
    As Alex grabs his running shoes, Bruce points out that no one can
    outrun a bear. “I don’t have to outrun him,” Alex replies. “I only
    have to outrun you.”‘

    It’s interesting the joke is referred to in 1988 as “recent.” I’d assumed it was folkloric.

18. Indeed it would, though if the two campers aren’t in Ursinia, then that’s neither here nor there. But I totally agree with you that we need to (somehow) figure out what’s a priori about the deontological constraints and why what’s not a priori isn’t a priori. Indeterminacy in the deontological constraints that we can argue for in a state of nature seems like the obvious answer to the second questions, and social rules (that take us of the state of nature) and joint commitments (that don’t) seem like the obvious ways in which those constraints could be made more determinate. Or anyway, this is what I argue in “The State of Nature” (2021).

19. Victor-I’ve always wondered whether your doing/allowing is quite the right way to capture this structure, though I will admit I can’t yet put my finger on what else we want to say. For one thing, I think D/A obscures two distinct issues. One-claims on agents—we are entitled to allow harms that we may not cause. (Though we may need to rescue when not too demanding). Two-it is worse to kill someone than to allow them to die. Admittedly, I am skeptical of this second formulation.

    I think two propositions are independently true, though Jump Behind makes their conjunction unseemly: (1) Archie has no duty to stay put, and (2) Archie is permitted to exploit the causal structure.

    So does Archie have a duty to rescue Bella? Assume Archie can just step out of the way (he doesn’t need to jump behind her), I think he may do so. She has no claim that he stay put. We see those sorts of cases all the time.

    Second, we can exploit causal mechanisms that we may not create. Assume A decides to fire indiscriminately throughout a room. B is very large and cannot move. I think it is perfectly permissible for C to jump behind B and use B as a shield provided that doesn’t change A’s pattern of fire. And, I think this would be permissible even if there was an exit door and C had equal chance of escaping death by running to the door or jumping behind B. Again, if B can suddenly move, C has no claim on B that B stay in place. To be sure, as with GP, if C can instead rescue them both, then C can’t just jump behind B.

    In Patrick’s Out of Control, the issue is that jumping behind Daphne redirects the car. Carol brings the threat to her. Because of the redirection, that wouldn’t be permissible.

    Now, Victor presses Guinea Pig. That’s a good push. Guinea Pig involves the failure to do one’s duty as the means by which the five are saved; whereas, in Room, the leg merely enables the rescue (it isn’t the failing to do one’s duty). But why then isn’t the answer that the duty to rescue is crucial here? You have to rescue the person in Guinea Pig and can’t point to the way the five will be saved to justify it; whereas, Archie gets to say that he has no duty to stay put to save her life.

20. Thanks Michael. That does seem a neat way to get beyond purely intuitive constraints. I’ll be sure to look at the paper.

21. Thanks Carl,

    I’m not sure that the equivalent of wrongly jumping behind happens almost never, at least if it is wrong to jump behind (giving the lie to my earlier response to James and PD that there are no implications in the real world of the right view about these cases). School bullies look for kids to beat up. Many people stay out of their way, hoping that they’ll pick on someone else. Or, no one has done their homework, teacher asks a question, pupils look at their desks hoping someone else will be asked. These have a ‘jump behind’ flavour to them, and we could have developed social norms that make us averse to avoidance in these cases. But we haven’t.

    There are also plenty of similar instances where jumping behind seems wrong – for example, where the person who will be beaten up or asked the difficult question about the homework is much weaker, more vulnerable, or systematical discriminated against. This will often be so in the real world, So don’t we have less reason to accept the idea that frequency of wrongdoing in cases like this explains our intuitions, making them unreliable at least as intuitions about basic features of morality?

    That being said, I’m open to the idea that there is no important difference between Jump Behind and Pull In Front. Almost everyone I’ve asked feels a strong intuitive difference, but it’s proved hard to explain it by pointing to anything that seems very significant, so perhaps we should just accept that our intuitions are unreliable in this case. That’s one response to the puzzle.

    Thanks Kim – yes, this would be an application of D-A that some find compelling in agent neutral cases – some think it wrong to turn a threat from one person to another other things equal, not the sense where we have prerogatives not to act, supporting allowing. Doesn’t your description of Guinea Pig beg the question? Whether we have a duty to rescue the one from the illness by giving them treatment depends on whether we should allow the one to die to save the five.

22. -Victor- Fair enough on GP. That was too quick. But as I’ve argued elsewhere (Means Principle and Optimific Wrongs)– Archie won’t have a duty to stay on the tracks even if there are five people behind him. They have no claim on him that he be the mechanism for their rescue. –I also argued, more controversially, that one may help Archie if he is sleeping on the track even if that does expose five people downstream of the one to the trolley. Or, imagine that the person trapped on the tracks asks you to give him a hand: I don’t think that there is any moral force to the five saying “no, leave him as the means by which we are saved.” So, in GP, I don’t think the existence of the five can generate a claim that the doctor not save the patient. Then, given the doctor’s baseline duty to rescue (imagining the five don’t exist), failing to rescue is failing to do one’s duty as the means by which the five are saved, and that’s the problem. (If I recall from when I presented my Means Principle & Optimific Wrongs at CELPA, you were skeptical…) But if I am right about propositions 1 and 2, isn’t Jump Behind permissible?

    -Michael- I love the Boorse and Sorensen article. Indeed, I think they go astray precisely because they overread the means principle so that exploiting causal structures violates it. (Something I argue in Acts, Omissions, and the Duty to Rescue)

23. Hi, Victor. On your final thought it’s because in our judgements we act with the knowledge we have or think we have. And, more generally, the concern is that the examples do not tell us anything interesting or morally significant etc about the real world.

    Assuming fing is using, there is more going on here than fing, the uncertainty and risk are of something else. So the dialogue between you and the objector does not quite capture what’s going on here.

    I suspect that one reason Patrick and others might disagree about Jump Behind is that he imputes using B as the reason for A’s action, while others may not, rather imputing getting away from the trolley as A’s reason for acting. We tend in these experiments too, to fill any potential gaps with applicable real world knowledge (of people, trains, bears etc). I tried to reformulate this avoiding the above, and I came to the conclusion that pulling is probably using anyway. So you can sketch it out as this.

    G: A will suffer harm unless A gs. If A gs, B will suffer harm.
    F: A will suffer harm unless A fs B. If A fs B, B will suffer harm.
    G behind: As in G, except that by A ging, B may be f’d.
    F in front: As in G behind, except that by ging, A f’s B.
    Gers: As in G behind, except that A and B may both G until one is eventually harmed. By Ging, the other may be f’d.

    This looks unlikely to tell us anything much. But I’m not convinced it’s a worse way of setting out these scenarios…

24. Thanks for the reply, Victor. I agree the bully and homework cases have a “jump behind” flavour, and may be quite common. But as they are also fairly trivial, there are no strong, universal norms condemning jumping behind, equivalent to those that condemn pulling someone to their death. My claim was specifically about the frequency of life-and-death situations of different kinds (“occasions for applying rules against jumping behind someone so they die rather than you, or running away from a bear that’s also chasing someone else, happen almost never”).

25. Victor: You’re right about the irrelevance of Bella having the option to pull Archie behind her to the permissibility of Archie jumping. But that’s the structure the case would need to have for someone to inevitably act wrongly by using isn’t it? If Bella has no choice, can she be said to ‘allow’ Archie to be used? Or to wrong him?

    But what do you think about my description of the case that it seems like if Archie jumps, he allows Bella to be harmed, but he does not merely allow her to be used. He uses her. So the doing / allowing distinction wrt harm points in one direction, but the d/a distinction wrt using points in the other. I suggest that’s partly why this is a tricky case to think through. We have the weaker constraint on harming, but the stronger constraint on using.

26. Hi James. I don’t think using Bella is or need be Archie’s motivation for jumping (he doesn’t jump in order to use her). But using her must be a condition for his jumping (he jumps because he will use her). Otherwise what would be the point in jumping? He’s just be a bit further away from the train and still be killed if Bella weren’t there.

    If he really jumps ‘just to get a bit further away from the train for now’ this looks impermissible, for he condemns someone else to death for this trivial reason. He has to jump behind only if it saves his life, and it will save his life only because Bella will stop the train.

    I think the reason I disagree with Victor and others is because (a) there’s a clash between our permission to avoid being used and the constraint on others which makes it hard to judge; (b) it looks like an allowing harm (it isn’t TOO different from jumping off the track) which makes it seem comparatively innocuous.

27. Pingback: Navigating the Labyrinth of Morality: PEA Soup Launches New Series on Philosophical Puzzles and Thought Experiments – Kitab Dar

28. Hi Patrick,
    Thanks for clearing this up. I think I agree with you on (b).
    I did not quite mean A’s motivation, which might well just be to save himself. Possibly, ‘he jumps because he will use her’ is a better way of putting what I think there may be disagreement about.

29. Kim, but if you accept that doctors are required to treat the one in GP, even though failing to do so will save five, it isn’t true that we are generally permitted to exploit the causal structure for our benefit. For by letting the one die, that is all the doctor would do. They would exploit the fact that the one will die for the sake of the five. Yet it is wrong to do so. So, why can’t Bella complain that it is also wrong for Archie to exploit the existing causal structure by jumping behind her in Jump Behind? If there is good answer to that, I think it must be that Archie will otherwise be useful in saving Bella. But now, as Patrick’s remarks suggest, we need a reason to prioritise Archie not being used over Bella not being used. And it’s hard to see what reason to give for this.

    One view is that Archie is not allowed to jump based on the idea that he would be using rather than allowing Bella to be used as Patrick suggests. That view might be reinforced by the intuition that it seems wrong for a neutral third party to help Archie jump behind.

    An alternative view that I find more intuitive, but less theoretically satisfying, is that Archie is permitted to use Bella to prevent himself from being harmfully useful on agent-relative grounds. On this view, we have a strong prerogative not to suffer harm by doing noting rbased on the fact that it would be useful to others that we don’t act.

    What then to say about Pull In Front, though? Well, in that case Archie does not merely use, he harms rather than allowing the harm, and that counts against pulling in front when compared with jumping behind. Again, I find that view intuitive, but not completely satisfying.

    So, overall, I still can’t see an easy solution to these cases that is both intuitive and theoretically satisfying. In the light of that it is tempting to explore debunking explanations, such as the one Carl suggests. But I wasn’t too sure how strong Carl’s debunking explanation really is once we recognise that a similar structure is involved in every day cases. Carl, you respond, that this is only true in cases that involve less serious harms. But I’m not sure how good that reply is; we seem to develop general intuitions about things like using, ducking, allowing to be useful, etc that are insensitive to whether the stakes are high or low. Our intuitions in the cases I set wouldn’t vary if we make the harms a broken finger rather than a loss of life I think.

    Anyway, that’s all to say that despite all your best efforts I’m still puzzled!

30. Jump Behind seems permissible. I’m inclined to think that a part of the explanation is that Archie does not infringe any right of Bella’s in the course of doing so, where as in Pull he does infringe a right to bodily autonomy (even when there aren’t trolleys involved, people have a right that we not lift them and move them, but lack a right that we not stand behind them). I wonder what others think about:

    *90 degree Turn*: Bella is safely to the side of the track; Archie is standing on the track, but facing her. He is also wearing some comedically long clown shoes, and Bella is standing on them. Archie can turn to face the oncoming trolley, moving Bella, who is on his shoes, in front of it, saving himself.
    I believe that turning is permissible.

    (Thompson said something related in her 1985 Trolley paper, when she tentatively suggested that it might be wrong to turn the trolley toward a bystander when the only way to do so was to use a nail-file that belonging to him. It is puzzling why Lilliputian property rights should make all the difference to permissibility when lives are at stake, but that is where my intuition goes here, I think.)

31. Thanks for the puzzle, Victor. I wonder what you think about the following line of thought (very cursory):

    In Jump, it seems permissible for A to jump aside, extricating himself from the harmful situation, by operating only on himsemf. The result is that Bella will be killed. The question is whether A permitted to jump behind in the second case. The complicating factor seems to be that by jumping behind, the relative placement between A and B is such that B will be used or useful for A’s survival. And some think that’s a weighty wrong-making featuer.

    Here is a line of thought that seeks to show that we do not have a (strong) reason to avoid placing ourselves in such a way that another will be used/useful for our survival, in situations relevant to Jump Behind.

    Jump-behind-then-aside: Suppose that, unlike in the original Jump, A cannot directly jump aside. Instead, he needs to first jump behind B, and only then is he able to jump aside. (Why? Perhaps the trolley track sits on a narrow bridge, such that directly jump aside would plunge A to certain death. So, in order to extricate himself from the harmful situation, A needs to first jump behind B, only then would there be firm ground for A to jump aside on).

    My sense is that if it is permissible for A to jump aside in Jump, then it is permissible for A to jump behind then aside in this case.

    Now consider the following: Once A performs the first jump (jump behind), is there a strong moral reason for A to also make the second jump (jump aside)? Or may A simply stay on the trolley track, perhaps just to conserve his energy. My sense is that A there is no strong reason for A to also make the second jump.

    One may then push back: But if A stays on the trolley tracak, then in fact B will be used/useful for A’s survival. That’s a weighty wrong-making feature, and it gives a strong reason for A to perform the second jump, which ensures that B’s death is in fact not useful to A’s survival.

    But I am inclined to think: Once A performs the first jump (jump behind), B’s death is ensured. It seems implausible to me to insist that, at the precise time the trolley collides with B, it matters morally, and in strong way, whether A is in fact standing *on* a piece of the trolley track or standing *beside* the trolley track. Certainly, from B’s perspective, I can detect no relevant complaint regarding where A will be standing.

    Indeed, suppose performing the second jump is moderately costly to A; it will break both of A’s legs. Again, I think it is implausible to insist that, because we have a strong moral reason to ensure that B’s death is not used/useful for A’s survival, A is morally required to perform the second jump.

    One way state my thought is that, in cases with the following features, there is no strong reason to ensure our own placement is such that another’s presence is not used/useful for our own survival: (a) the initial placements seem arbitrary to begin with, and (2) agents can change their initial placement by operating only on themselves.

    The argument for this thought is that, if there were such a strong reason, then A would be required to perform the second jump in Jump-behind-then-aside. But that seems implausible to me. So, A is permitted to jump behind B in Jump Behind.

32. My daughter just passed her driving test, so it may well be that scenarios like these will become increasingly common!

    T, isn’t your appeal to rights, like Kim’s appeal to duties, just begging the question? You say that none of Bella’s rights are infringed in Jump Behind, but surely Archie would violate Bella’s right to be protected from the trolley if he would stop the trolley by staying where he is and he would not be harmed, or would be harmed to a trivial degree. This suggests that whether he is violating her right depends on whether he is permitted to jump behind her. If he violates, he infringes, so we cannot appeal to infringement without begging the question. Turning doesn’t seem permissible to me in your clown shoes case. That case doesn’t seem importantly different to Pull. What is the difference for you?

    Erik, this is a nice line of argument, but I wonder if your Jump Behind then Aside has a secondary permissibility flavour about it. It is permissible Archie to jump behind in that case because he could permissibly save his life without using Bella by jumping behind and then aside, and jumping behind but staying is a Pareto improvement over this permissible alternative. I think we shouldn’t draw conclusions from that case for the case where there is no possibility of surviving without using.

    Compare the standard secondary permissibility case (I think it’s Frances Kamm’s) where I can either turn a trolley from five to one, or throw the very same one in front of the trolley to save the very same five. However, suppose that turning is worse for the one and the five than throwing. It is then permissible to throw because throwing is a Pareto improvement over turning, and turning is permissible becuase it doesn’t involve using. We shouldn’t conclude from this that it is permissible to throw where there is no turning alternative.

33. Thanks for the reply Victor! I agree with the astute observation that my case does have a secondary permissibility flavour it. One thing to say in response, however, is this:

    In Kamm’s original discussion of secondary permissibility (if I recall from my graduate seminar with Johann Frick), the principle of SP applies in cases in which (a) it is already morally permissibility to impose some degree of harm H on A in order to achieve some good, and (b) there is an alternative that imposes *less* harm than H on A that achieves the same good. So, the illustrative example I was given as a graduate student: If it is permissible to turn the trolley onto the One, thereby killing him, in order to save the Five, then it would also be permissible to push the One onto the trolley track, thereby paralying him, in order to save the Five.

    I take it that, as far as secondary permissibility is concerned, we should *not* be netural between: (1) turing the trollye onto the One, thereby killing him, in order to save Five and (2) throwing the One off a footbridge, thereby killing him, in order to save Five. The reasons against the latter is strong, even under SP.

    In my case (Jump-behind-then-aside), however, the harm to B is the same, whether I jump-behind-then-aside, or I jump-behind-but-stay-on-the-track. In other words, the alternative options A has involve the *same* harm to B. Insofar as you share my intuition that it is permissible for A perform only the first jump, even if performing the second jump involves no extra cost to A, I do wonder whether the Principle of Secondary Permissibility as Kamm formulates it, is the correct diagnosis here.

    Of course, you might think either that Kamm’s formulation of SP is too restrive, or that once the detailed are spelled out, there is greater reason for A to perform the second jump than staying put.

34. I’m not sure Erik; I’d need to think more about SP, but I do tend to think that the formulation you outline from Frances is too restrictive (it’s been a long time since I’ve read her work on this, so I’ll have to trust you on the details). I think that there really isn’t a reason against throwing a person in front of the trolley when the alternative is to turn it onto that person. One way of understanding this is that the wrongness of harmfully using a person depends on some baseline that determines whether the person is harmed, and in SP cases, the person is not harmed in the relevant sense, for they are not worse off than they would be in the permissible alternative. At any rate, the fact that there is a plausible SP explanation of Jump Behind then Aside I think suggests that it’s unsafe to draw conclusions from that case for Jump Behind.

35. Thanks to everyone for their thoughts about the puzzle. As we said, anyone is welcome to propose a puzzle to discuss on PEA Soup. We might even solve some of them!!!

Comments are closed.