Print This Post Print This Post
The Loop Case Poll
By S. Matthew Liao

Here’s another case on which you can vote. It came from Judith Jarvis Thomson’s 1985 article “The Trolley Problem.”

The Loop Case: A trolley is headed toward five people, and it can be redirected onto another track where one innocent bystander sits. However, the track loops back toward the five. Hence, if it were not the case that the trolley would hit the one and grind to a halt, the trolley would go around and kill the five. Assume also that if five were not present, the trolley would not loop toward the one, but would continue harmlessly down the track (Kamm, p. 92).

By the way, all the cases polled will be kept in an archive, and you can view their results from the Polls navigation tab at the top of any page. You can also continue to vote on previous cases from a case’s respective post. For example, for the Munitions Grief Case, see the Kamm Poll Post.

Finally, thoughts and comments on the Loop Case are welcome here.

Happy Voting!

The Loop Case: Is it permissible to redirect the trolley towards the one? [See The Loop Case Poll Post for details]

  • Yes (66%)
  • No (29%)
  • Not sure (5%)

Total Votes: 79


Loading ... Loading ...


  1. 1. Posted by Mike Otsuka | August 12, 2007 1:07 pm

    Unlike some other examples in normative ethics such as those involving terror bombing versus strategic bombing, the Loop Case is not modelled on any real-world cases with which we are familiar and about which we have already formed reactions. The first encounter of readers of this blog with this particular problem will probably have been either in the pages of Thomson’s ‘The Trolley Problem’ to which Matthew refers or through discussion of this case that was ultimately inspired by this particular article. In her inauguration of the Loop Case, Thomson drew our attention to the similarities between it and the original Trolley Case and persuasively argued that diversion in the Loop Case is permissible just as it is permissible in the original Trolley Case. It is an interesting question whether the (as of the time of the posting of this comment) high proportion of ‘Yes’ votes in Matthew’s poll is an artefact of the fact that our initial exposure to this case has been so strongly influenced by the manner in which Thomson introduced this case to us in the context of the original Trolley Case. Note that, by contrast, the opinion of those who took Marc Hauser’s online Moral Sense Test was divided 50-50% regarding the permissibility of diverting the trolley in a version of a looping case highly similar to Thomson’s. Presumably many of these respondents were internet users who had not been previously exposed to Thomson-influenced discussion of the trolley problem. Moreover, Hauser’s statistics were drawn only from people’s responses to the first case that they encountered in the online survey, thereby screening out the influence of other cases on their convictions regarding permissibility.

  2. 2. Posted by Jeff Huggins | August 12, 2007 3:50 pm

    Mike, thanks for the very helpful post. Given your knowledge of the use of these types of case polls, I’d like to ask a few questions that perhaps you (or others reading this) could answer:

    Is anyone doing research with these case questions (especially the more straightforward ones and simple variations) that is robust and statistically significant and that meets other criteria for scientific research? Is Marc Hauser’s online voluntary poll the closest to meeting most of these criteria, (or at least some of them), or is someone else doing research on reasonably large numbers of people in a more robust way?

    I’m particularly interested in two aspects of the above questions: 1) Whether someone is doing top-quality research with a specific segment of individuals (for example, American college students who have previously not been exposed to these cases or to moral philosophy), to at least get accurate trustable answers representing the intuitions of that group; and 2) Whether someone is doing top-quality research of like groups across diverse cultures, in order to understand how ‘universal’ the various intuitive answers are?

    I’ve read Hauser’s book, Moral Minds, although I can’t remember the details he shares about his methodologies, sample sizes, and so forth (I’ll look back at it this coming week). And, I’ve taken the online poll (Moral Sense Test). That said, it surprises me, a bit, that so much focus is being placed on intuitive answers in books such as Kamm’s, at such detailed and intricate levels, based largely (it seems) on individual/personal intuitive answers or very small informal samples, not on research that meets the normal criteria for excellent research.

  3. 3. Posted by Guy Kahane | August 12, 2007 3:58 pm

    Mike, you offer a possible (and plausible) causal explanation for the difference in results. But I think you also hint that Hauser’s results have greater epistemic standing. But why should this be so? Why should prior exposure to a similar but structurally simpler case bias rather than improve our intuitive assessment of the Loop Case? (In fact I’m not even sure if Hauser’s measure is in any interesting sense even PSYCHOLOGICALLY purer.)

    To make my point more vivid, consider Moral Mary. Moral Mary has led a very secluded life. She was exposed to just the minimum amount of moral dilemmas that is required to obtain proper grasp of moral concepts. If we ask Moral Mary for her intuitions about the Loop Case or self defense or capital punishment or torture, do her replies have higher rather than lower epistemic standing?

    (Of course, you may have in mind a more specific debunking causal explanation of the influence of being first exposed to the Trolley Case or Thomson’s presentation. I’m just arguing that this causal influence has no debunking force on its own.)

  4. 4. Posted by Mike Otsuka | August 12, 2007 10:25 pm

    Jeff: I’m not aware of any research on intuitions about trolley cases that would meet the standards of a scientifically representative survey. The only surveys of which I’m aware are either (i) self-selecting internet polls such as Hauser’s Moral Sense Test, a BBC poll, and of course Matthew’s, or else (ii) Joshua Greene’s experiments involving university students who have volunteered to have their brains monitored by fMRI scanners to see what’s going on inside their heads when they react to different cases. Hauser has a pretty large international sample of data from those who speak Spanish and Chinese as well as English. He notes that, among this international group of people who surf the internet, there is very little cultural variation in responses. I also recall that he mentions, in his book, that he has plans to survey isolated tribes of hunter-gatherers to see whether their reactions vary from the reactions of internet users.

    Guy: I agree that having previously reflected on another similar case might – for the reason you offer – make one’s intuitive judgment about a given case more rather than less reliable. So I wouldn’t want to advance the general claim that intuitions about cases considered in isolation from other cases are more reliable. A problem with the Loop Case, however, is that our intuitions regarding the permissibility of diverting in that case seem particularly vulnerable to the context in which it’s presented. Suppose, for example, that one is first presented with the Bridge Case in which there is a ‘person [who is] is on a [pedestrian] bridge over the track and cannot move off it. If we move a pole, then it will topple him gently into the [path of the] trolley that is headed toward five people; his being hit will stop that trolley and kill him’ (quoting from p. 143 of Kamm’s book). Most intuit that toppling the one is impermissible. Now suppose one is next presented with Fischer’s Ramp Case in which one can stop the trolley from hitting the five by opening a draw bridge. So doing will send the trolley up the ramp of the open draw bridge and into the air, where it will hit one person standing on a pedestrian bridge overhead. The trolley’s hitting the one will stop its forward progression and cause it to fall into the water below the draw bridge. The trolley’s hitting the one will also kill him. If, however, the one had not been overhead, then the trolley would have managed to fly over the gap, land on the downward sloping ramp, and hit and kill the five. I think that many will assimilate this case to the Bridge Case and intuit that it is impermissible to send the trolley up the ramp and into the one. But the Ramp Case is simply a vertical version of the Loop Case. Suppose we assume that non-isolated intuitions about the Loop/Ramp Cases are more reliable than isolated intuitions about these cases. Now the following vexing question arises: which of two conflicting non-isolated intuitions about the Loop/Ramp Cases is more reliable — the intuition of permissibility that was formed in the context of the Trolley Case, or the intuition of impermissiblity that was formed in the context of the Bridge Case?

  5. 5. Posted by Dale Dorsey | August 13, 2007 4:55 pm

    Hi all. This is an interesting post, and I think Michael makes a very interesting point about the Loop case in particular in light of the draw bridge case. I’m very sympathetic to the assimilation of the Loop case to the Bridge case; but it seems to me this is a feather in the cap of the (broadly speaking) consequentialist. In particular (Michael will be having some deja vu, as I’ve sent him this argument before; I proceed despite his elegant refutation), it seems to me that assimilating the Loop case to the Bridge case ought not to cause us to rethink the Loop case, but rather the Bridge case. This is probably just table-pounding, but let me state a different case that might motivate my point. Call this “two levers”.

    Imagine that you’re the bystander in Bystander at the Switch. This time the levers are inside a little “conductor’s hut” or something. But as you run in there to turn the Trolley, you find not one lever but two levers, corresponding to two points at which the track could diverge. The first lever turns the track at only the first point, before the five, and onto the track with the one. The second lever turns the track at both the first point and the second point. The second point is located behind the one, sending the trolley back to the main track, before the five. So, effectively, pulling the first lever turns the case into a normal Trolley. The second lever turns it into a Loop case (the one will slow the trolley down, etc., etc.). Now it seems to me that there is simply no reason at all to say that either lever makes a moral difference. In other words, it just doesn’t seem to me to matter morally to pull the first lever, as opposed to the second lever. Furthermore, if there is even just a *fraction* of benefit to be gained by pulling the second lever (imagine, for instance, that getting the Trolley back to the main track would allow people to get to work on time in a way that they wouldn’t have gotten to work a few minutes late had you pulled the first lever), there is a substantial–it seems to me overwhelming–moral reason to pull the second lever.

    Now we all agree that pulling the lever in Trolley is ok. I’ve tried to argue above that Trolley and Loop should be assimilated. But if that’s right, why aren’t we rethinking the Bridge case if the Loop case is morally assimilated to it? Perhaps the above intuition is colored by some deep welfarism or consequentialism or something, but I find it hard to motivate the thought that there is a reason not to pull the second lever, or if there is that it wouldn’t be overridden by even the slightest benefit. But I leave open the possibility that my intuitions here are idiosyncratic or colored by presentation effects.

  6. 6. Posted by S. Matthew Liao | August 13, 2007 6:18 pm

    Dale, thanks for your “two levers” case. Suppose the one person were not on the side track. We would still pull the first lever, because then the trolley would hit no one. But, so some might think (Kamm’s DTE notwithstanding), it would make no sense to pull the second lever, because the trolley would just loop around and still hit the five. If so, that might suggest to some that there are some differences between the two levers. In any case, isn’t the “two levers” case just a redescription of the original trolley and loop cases? If so, it seems that one would say regarding the “two levers” case whatever one would have said regarding those two cases, or, so I would have thought.

  7. 7. Posted by Dale Dorsey | August 13, 2007 6:30 pm

    Matthew –

    The one difference in my case is that the protagonist has the option to *make* it into one version or the other. Now you’re probably right that if you have a “no loop” intuition then you would have a “first lever only” intuition. But (maybe this is just me?) I think it helps to bring out the moral equivalence of the cases if you consider whether it is morally impermissible to *make it* into a loop case, versus *making it* into a standard trolley case. In the usual presentation, you’re just hit with a case you can’t change. But if you can select the case, my intuition tells me, it doesn’t make a lick of moral difference which lever you pick. Either way, the one is dead, the five are saved. But I think the more telling intuition is that, even if you have an intuition in favor of “first lever only”, it is not impermissible to pull the second lever if this would yield additional benefits (say, fewer headaches because of lateness). I suspect I’m pounding the table but I suppose I’m exploring intuitions that I hold strongly.

    And you’re right that if there were no one on the side track, it would be absurd to pull the second lever. But, it seems to me, in the case in which there *is* one on the side track, it makes very little difference whether you make this into a loop case or a standard trolley.

  8. 8. Posted by David Wasserman | August 14, 2007 3:01 pm

    I think there are two ways of blocking an inference from the permissability of choosing the loop over the diversion in Two Levers for any benefit to the accepability of the original Loop. The first way involves, I think, Kamm’s principle of secondary permissibility. If the agent can harm someone permissibly by one means (first lever), she may cause the same person equal or less harm by an otherwise-impermissible means (second lever) if there’s some benefit in doing so. The second reason is that in the context of cases like Two Levers, the otherwise-impermissible action no long involves using-as-a necessary-means — in the context of the agent’s choice, the one’s being hit is not necessary to save the five, since the agent can pull the first lever instead. True, the one’s being hit IS a necessary means to whatever marginal benefit accrues from pulling the second lever, a benefit Kamm might regard as an irrelevant utility, and that feature might casue some to reject the choice of the second lever except for major benefits.

    I should also note that we don’t all agree that’s it’s OK to pull lever one — as I said in discussing Chapter 4, I’d assimilate the original trolley case to standard cases of killing a smaller number to avoid killing a larger number.

  9. 9. Posted by Jeff Huggins | August 14, 2007 5:25 pm

    Guy and David and others,

    Guy writes: “Why should prior exposure to a similar but structurally simpler case bias rather than improve our intuitive assessment of the Loop Case? (In fact I’m not even sure if Hauser’s measure is in any interesting sense even PSYCHOLOGICALLY purer.) To make my point more vivid, consider Moral Mary. Moral Mary has led a very secluded life. She was exposed to just the minimum amount of moral dilemmas that is required to obtain proper grasp of moral concepts. If we ask Moral Mary for her intuitions about the Loop Case or self defense or capital punishment or torture, do her replies have higher rather than lower epistemic standing?”

    David also seems to point out that intuitive answers sometimes differ, even among moral philosophers, and even regarding what are considered to be the more straightforward cases.

    One issue, I think, is the binary classification implied by ‘permissible’ and ‘not permissible.’

    In responding to moral dilemmas, of course, there are often a range of choices, and they may cover a wide range of moral value, i.e., as represented here by a five-point scale:

    (1) The ideal choice (the best, as in most moral, solution);
    (2) A reasonably good and understandable choice (perhaps not ‘ideal’ but still falling well within a range of moral permissibility);
    (3) A not-so-good but still permissible choice;
    (4) A choice that is perhaps understandable (under the difficult circumstances of the situation) but is not to be celebrated or taught;
    (5) A choice that is clearly not permissible and that perhaps calls for some remedial consequences (depending on the violation, the situation, and the person’s mental state, etc.).

    If a moral case is presented to informed moral philosophers, (let’s assume it offers five possible choices), and if the bulk of the responses are spread significantly among three of those choices, then, presumably, those three choices would at least qualify as ‘understandable’ and ‘reasonable’, even if two of them are not ‘ideal.’ More than one may be permissible. How then shall we eventually decide which of the three choices is the ‘ideal’ (1) and which other choices will be deemed (2), (3), or (4)?

    If intuition-based research with moral philosophers (or larger-scale research with the broader population) yields a substantial number of positive responses to more than one option in a moral case, and especially if the answers are not strongly in favor of one option over others, then, clearly, it is difficult to isolate the ‘ideal’ answer from the others without appeal to some other source of moral understanding. Those other sources should involve, in my view, an understanding of the foundational ‘effective’ function of morality itself, an assessment of considerations such as ‘what if everyone made this choice in this type of situation?’, an understanding of human abilities and limits, and an understanding of the role of ‘a basic sense of stability and security’ in the overall functioning of human life.

    In response to Guy’s point (and questions) regarding ‘Moral Mary’, I offer four thoughts for consideration: One is that Moral Mary’s initial intuitive response to a given moral case could, of course, fall anywhere within the range (1) to (5), just as a moral philosopher’s initial intuitive response could fall anywhere within that range.

    The second involves a comment made by Thomas Jefferson in a letter to a nephew, in which Jefferson wrote: “State a moral case to a ploughman and a professor. The former will decide it as well, & often better than the latter, because he has not been led astray by artificial rules.” (ref. Jefferson’s letter to Peter Carr, August 10, 1787, written from Paris).

    The third is, of course, that a person’s answers to certain types of questions depend (to a degree) on how a question is framed and also on the person’s prior experience and immediate frame of mind, or frame of reference. (Economists now realize this, and people who do lots of research with humans also experience this.)

    The fourth is that, in my experience, people sometimes (perhaps often) flip-flop their intuitive answers among more than one plausible (to them) solution. In many of these situations, it is unclear whether their answers are getting better (or at least more informed) or worse. As I’m sure we’ve all experienced, people can go from answer ‘A’ to answer ‘B’ and then back again to answer ‘A’, and then to ‘B’ again, especially in the more complex cases. That’s one reason why, in my view, we must ultimately also appeal to sources of moral understanding other than the intuitive responses in order to ultimately differentiate solutions (1), (2), (3), and etc. as identified above.

Post a comment

Name: (required)

Email Address: (required) (will not be published)