Trolley Problems

by John Holbo on September 25, 2010

Evergreen philosophy topic! Sure to inspire much friendly discussion!

I don’t usually lecture about the stuff myself, but this semester I decided to, so I cartooned up some images for the PPT slides. So the first thing I have to say is that if anyone has a use for ‘em, I’ve released ‘em under a CC license.

Click for larger, for all these.

Yer basic Cartoon Modern trolley. (I thought of going more for a “Toonerville Trolley” look. But it’s not my style.)

basic trolley scenario

The Classic Trolley Situation

bridge situation

The Fat Man On The Bridge. (Adapted from a Robert Ludlum novel, The Adipose Conundrum.)

Now, of course, there have been many fine parodies and mockeries of Trolley Problems. Personally, I think rewriting Green Eggs and Ham might be a good approach. “Would you like the greatest Good?” “No, I do not think I should!” Then foxes and boxes and trains and etc. (Have a go in comments.)

I also like the fact that Wikipedia flags a relatively flat, straight, standard statement of the problem as containing ‘weasel words’.

A semi-serious issue: Trolley Car Ethics is often singled out as the apotheosis of analytic-style absurdity. Anglo-American-style philosophers are forever accused of suffering from a severe, Aspergers-like condition, causing them to be incapable of apprehending anything but the model train sets in their basements. (I’m told David Lewis had elaborate model train sets.) But it seems to me the criticism is more effective if made more mildly: the problem is not that Anglo-American, analytic-style philosophers (call them what you like) are severely autistic but that they are mildly whimsical. This style of example-mongering complements a characteristic style of seminar room cleverness. There’s an intellectual ethos that goes with this social ethos; and, if you ask me, there’s a lot to be said for it. But the concern is that it’s going to clog the so-called ‘intuition pump’. People aren’t reporting their ‘intuitions’. Period. Rather, they are telling you what they would say, once you’ve set the mood for a good, seminar-style go-round. The examples are supposed to take you out of the seminar room, imaginatively. But plausibly they do the opposite – are really designed to do the opposite. They export the socially and intellectually distinctive seminar room atmosphere to a whole, slightly off-kilter world.

I decided to teach the Trolley stuff because I recommended that my students listen to one of Sandel’s “Justice” lectures, and he talks about Trolley stuff. I’m pleased with how the cases play in the classroom, as it turns out. Good discussion and heightened appreciation of general issues. Utilitarianism and its discontents. I also think the bug – namely, these situations are silly – can be a feature. You pump students for their ‘intutions’. Then you talk about how making the stories silly is likely to distort responses. What ‘normal’ elements of response are amplified or suppressed? Which in turn is a good lead-in to further thoughts about utilitarianism and its discontents.



Adam 09.25.10 at 8:08 am


Timothy Scriven 09.25.10 at 9:18 am

I’ve been thinking about creating a more emotionally charged series of thought experiments with the same basic structure, and comparing responses to these with responses to whimsical trolley problem style scenarios for a while now.


Henri Vieuxtemps 09.25.10 at 9:40 am

Use children and puppies.


John Holbo 09.25.10 at 9:52 am

Yes, the comparison would be worthwhile. Children and puppies are fine. Maybe episodes of “The Wire”.

One thing I probably should have mentioned in the post, is that one reason these Trolley cases work in the classroom is that, in a way, they do double-duty as ‘what is utilitarianism’ plus ‘how to advance a thesis and make objections in a philosophy seminar’ lessons. But, by the same token, they don’t do great primary duty as ‘how you think about right and wrong, whether you are in a philosophy seminar or not’ measuring instruments.


John Holbo 09.25.10 at 10:16 am

Just to make one rather unimportant clarification about my personal experience: my actual trolley car lecture bits this semester are yet to occur, so it could all still disappoint. But preliminary student forum discussion has been good, and in the past I’ve co-taught with someone who did Trolley stuff and I was surprised with how well it played in the discussion sections I lead.


Tom Hurka 09.25.10 at 12:15 pm

It would help to understand what the point of the trolley case is. It’s not at all about “utilitarianism and its discontents.”

Let’s say you start out persuaded that utilitarianism is false, because it would be wrong to e.g. frame and execute one innocent person in order to prevent a riot that will kill five. Then you ask *why* that would be wrong, and a natural answer is that you’d be killing the one as against letting the five die: it’s morally worse to actively cause harm than just to allow it to happen, and you’d be actively killing the one.

But the trolley case shows that that can’t be the right answer. Even non-utilitarians (or most of them) think it’s *not* wrong to turn the trolley, even though that too involves actively killing one to save five. So the reason it’s wrong to frame and execute the one in the riot case can’t be just that it’s always worse to kill than to allow to die — the trolley case shows that it isn’t.

Of course if you start out with an economist’s attachment to utilitarianism you won’t be interested in why utilitarianism is false. But, amazingly, not everyone is a utilitarian. And they may want to understand what general principles underlie their anti-utilitarian intuitions. The trolley case is then a *problem* for them, because it shows that a simple answer won’t do. The case actually counts in favour of utilitarianism, by showing that the alternative will have to be much more complex than at first appeared.


Alejandro 09.25.10 at 12:33 pm

I’m told David Lewis had elaborate model train sets.

Surely you mean elaborate modal train sets.


John Holbo 09.25.10 at 1:15 pm

Tom, I don’t really think I am confused about the basic uses to which people think the Trolley Case(s) can be put. It is certainly useful for introducing consequentialism. Most people, in my experience, are willing to throw the switch in the basic case. 5 > 1. That’s consequentialism. It’s good to point out to people that they are on board with a lot, saying yes to that. But then they don’t want to push the fat man. Well, then: why not? I don’t think it’s so obvious that this further data point counts in favor of utilitarianism, as you say. Or against. It’s more of a ‘one man’s modus ponens is another man’s modus tollens’, no? A lot of this has to do with trying to put our finger on the character and source of our discontent with utilitarianism. Which is difficult. Hence ‘utilitarianism and its discontents’. Is there some reason why ‘discontents’ is a bad word for this, in your book? (It may be that you are reading it as ‘utilitarianism and what is wrong with it’. But that’s not it. It’s more like ‘civilization and its discontents’. I don’t intend a Freudian analogy in any strong sense, but I did mean to indicate that sort of tension.)


Juan Comesaña 09.25.10 at 3:45 pm

I think that Tom Hurka is right. Utilitarianism is a red herring (it gets more complicated with Consequentialism). Start with the transplant case, where no one would kill the one to provide organs for the five. We are already in non-utilitarian space. So now you ask, why? Obvious answer: killing vs. letting die. Enter the trolley case. Killing vs. letting die can’t be the answer. It’s not that you cannot use a trolley case to introduce utilitarianism (nothing wrong with that), but rather that the main purpose of trolley cases in the literature (Foot, Thompson, etc.) had little to do with it.


y81 09.25.10 at 3:55 pm

Do most people (e.g., most students in a seminar) vote for throwing the switch? That surprises me. I am probably on board with [what some would argue are] real-life equivalents of throwing the switch (like having wars, or allowing coal mining), but I don’t think I would vote in favor of throwing the switch.

I assume that, if you won’t throw the switch, then a fortiori you won’t push the fat man.


Juan Comesaña 09.25.10 at 4:03 pm

(Awesome cartoons, by the way.)


John Holbo 09.25.10 at 4:18 pm

It’s odd how prevalent the ‘throw the switch but not push the fat man’ response is. And students perfectly well see the oddity of what they are saying. They see that they are inclined to distinguish two situations concerning which they can’t really point to a very convincing moral difference. They are utilitarians with reservations, but they can’t articulate the nature of their reservations in a plausible way. (I guess this is Tom’s point.)


geo 09.25.10 at 4:21 pm

That’s consequentialism

If it’s not too tedious, would you explain the difference between “consequentialism” and “utilitarianism”?


Chris Bertram 09.25.10 at 4:32 pm

Geo: utilitarianism is a sub-type of consequentialism. Consequentialism says: bring about the best consequences; utilitarianism includes a view that when we compare states of the world (sets of consequences) we should score them in terms of the amount of utility (net welfare) they contain. A view that said that you should maximize the amount of aesthetic value in the world would be consequentialist, but it would not be utilitarian.


Norwegian Guy 09.25.10 at 4:36 pm

It’s odd how prevalent the ‘throw the switch but not push the fat man’ response is.

I would guess one difference between the switch and fat man cases is that while the switch can’t throw itself, you don’t need to push the fat man. He can jump himself.


geo 09.25.10 at 4:57 pm

Thanks, Chris. But why isn’t aesthetic enjoyment a utility? Surely beauty enhances our welfare?


ebenezer smooth 09.25.10 at 5:01 pm

The difference is proximity. Pushing is physical contact.
We find various ways to avoid the full burden of responsibility. It moves up the line from pusing to flicking the switch to agreeing that someone should do it etc.
The military is utilitarian and run on forced separation of commander and commanded. If the military is your model for society then you’re fine.


nnyhav 09.25.10 at 5:04 pm

Simplest version of @1, despite policy of “do not feed the trolleys”,

I was walking with the Ethicist across the footbridge over the tracks leading to the switching yard when he stopped and [interpose canonical statement of problem]. So I threw his fat ass over the railing into the path of an oncoming train.


Alex 09.25.10 at 5:12 pm

However, the Milgram experiment with the fake electric shocks has been replicated in setups where the experimental subject had to physically force the actor’s hand onto the contacts, without substantially different results.


John Casey 09.25.10 at 5:44 pm

I’m curious about how far from actual physical contact we can get, and whether changing that parameter would elicit a different response. I.e.,

1. Push the fat man
2. As the trolley driver, you switch the car onto the track with 1 victim from the track with multiple victims.
3. As the trolley driver, you do nothing to switch the car, which is already directed towards the 1 victim.
4. You are the president of the trolley company, fortuitously riding on the death trolley. You direct the driver to switch to the 1 victim track.
5. You are the general counsel for the trolley company. You write and publish a policy directive stating that where there is a choice, death trolleys should be switched onto the track with the fewest victims.

In at least some sense, each of these is morally equivalent. Yet they will almost certainly not elicit a uniform response, yes?


Borderline Troll 09.25.10 at 6:30 pm

Your “classic trolley situation” isn’t classic at all. You have depicted real humans on the main track, but on the shunt track you’ve put a ginger. Gingers, as is well known, have no souls, and therefore causing them to die does not pose any serious ethical problem.


Borderline Troll 09.25.10 at 6:36 pm

Okay, okay, if I’m gonna comment as “Borderline Troll” I guess I should add something more productive; in all seriousness, the race of potential trolley victims has been shown to significantly influence ethical intuitions in surprising ways.


Henri Vieuxtemps 09.25.10 at 6:47 pm

What does the legal system have to say about it? Suppose there is no negligence whatsoever on the part of the company; the brakes fail by a freak accident. Does switching to the other track and killing one person there make the driver/company liable?


Kragen Javier Sitaker 09.25.10 at 7:08 pm

I am not an expert on Wikipedia, but I suspect the weasel words are the following:

some non-utilitarians may also accept the view…. Opponents might assert that, since moral wrongs are already in place… Additionally, opponents may point to the incommensurability of human lives.

It might also be justifiable…

My inner Wikipedian is crying out, “Which non-utilitarians? Do they accept the view or don’t they? Are these opponents who might advocate non-participation purely imaginary, or are they real? Who are they? Do people actually point to the incommensurability of human lives, or is that a hypothetical strawman being used to attack hypothetical incommensurabilists? Anything might be justifiable; is there a school of thought that argues that the following ellipsis is justifiable?”

Unfortunately I don’t have enough of a background in ethical philosophy to answer these questions myself, or I’d fix the article and remove the “weasel words” tag. However, I assert that anyone who reads this article and has the relevant knowledge has a utilitarian obligation to edit the article so that the trolley runs over and kills the weasel words.

Wikipedia, not being written in the voice of any particular author, has little tolerance for descriptions of what hypothetical people might believe as logical consequences of premises they hypothetically hold, because that style of discourse has far too much scope for rhetorical abuse in the hands of effectively anonymous authors discussing contentious matters such as Scientology, the Armenian Genocide, or the Gaza Strip. There are an infinite number of plausible chains of reasoning by which one can plausibly derive odious moral consequences from the stated beliefs of one’s ideological opponents. Consequently this sort of thing is frowned upon, because it tends to lead to unproductive edit wars.

However, it would be entirely justifiable to say, “Thomson points out that a hypothetical utilitarian could reason that taking an action in the situation assumes responsibility for the consequences,” assuming she does, of course. The issue is the use of the “some say” device (in this case, worse: “some could conceivably say”) used by unethical journalists to inject their own opinions into ostensibly factual articles, cloaked in a thin veil of objectivity. I don’t think it’s being used that way here, but some clarification would help.


dsquared 09.25.10 at 7:15 pm

#17: I seem to remember that the face-to-face versions of the Milgram experiment did have a lot less compliance, but still a long way from zero. If I was downstairs or my copy of “Obedience to Authority” was upstairs, I’d check.


bianca steele 09.25.10 at 7:18 pm

@20 Good point, but from what I understand of Wikipedia’s editing policies: If the article cited in the footnote says “some say,” the Wikipedia article should do the same. Otherwise, the editors are publishing their own original research, which is explicitly against the rules.


bianca steele 09.25.10 at 7:22 pm

Re. the trolley problem: I don’t see why a non-major, encountering this problem, would not conclude that the philosophical way to approach a utilitarian-style dilemma is to couch it in terms of the trolley problem, especially if they’ve never run into the seminar style John Holbo describes above. Most people have not studied philosophy and really dislike hair-splitting forms of argument. The cleverer you get, the more they are certain you are missing the point. And they’re not entirely wrong. For example, it seems pretty obvious from many discussions I’ve seen that certain answers are just not in play; you can’t say, “but I think it makes a difference that the trolley is said to be ‘out of control,’” because it isn’t part of the essential nature of the issue.


Lizardbreath 09.25.10 at 7:35 pm

I’d strongly surmise that almost all of the ‘would throw the switch, wouldn’t push the fat man’ answers can be accounted for by not being able to accept the premise of the question, that pushing the fat man would stop the trolley. If you’re asking students for their intuitions, you’re asking them to make a general assessment of everything they know and feel about the situation, and expecting them to be able to discard their sense of the probable and accept both that the fat man’s body would stop a trolley that would be unimpeded enough by four (presumably leaner) bodies that it would still kill the fifth person, and that the student would be certain enough of this to act immediately on that basis. Even though that’s the stated premise of the question, it seems implausible enough to me that it’s going to throw off answers: I would (I think) flip the switch in a real-life situation; I would not (I know) push a fat man in a real life situation, because there is no remotely plausible real world situation because there is no chance I’d think it would work; and so I’m not sure what I’d do if I knew pushing the fat man would work, because I can’t really conceive of having that knowledge for real.

Has anyone tried coming up with a less whimsical and more plausible fat man analogue, and seeing if the plausibility brings people’s intuition in line with the switch problem?


Kragen Javier Sitaker 09.25.10 at 7:39 pm

@22: Bianca, there are cases where Wikipedia policies conflict with each other, but I don’t think this is one of them. If it’s Philippa Foot saying that “some say”, that attribution should be clear from the article, at least in the form of a footnote, and it isn’t clear. Given that Foot may herself be a somewhat unreliable narrator, since she was writing an article about a thought experiment which could legitimately include imaginary philosophers with various points of view, I think it would be better to footnote the people who actually said the things that “some say” rather than Foot’s secondhand assertion (even though Wikipedia prefers to depend on secondary sources rather than primary sources, so a philosophy textbook or a review article would be even better).

I’ve added a slightly edited version of my comment on the article’s discussion page, in case this comment thread doesn’t result in some bystander throwing a fat footnote in the path of the weasel words careening out of control.


Aulus Gellius 09.25.10 at 7:52 pm

geo @14:
Aesthetic value can be a utility. But a utilitarian, though he’d presumable prefer “more aesthetic value” to “less aesthetic value,” would, in at least some circumstances, be willing to sacrifice aesthetic value for other goods. Chris Bertram’s hypothetical aesthetic-value-maximizer would be willing to sacrifice a little aesthetic value for a lot of aesthetic value, but would never sacrifice even a little bit of aesthetic value for the sake of, say, someone’s health or personal fulfillment.
And consequentialism, at least as I understand it, doesn’t even have to be that close to utilitarianism: a consequentialist could be in favor, not of maximizing anything, but of maintaining some sort of balance between, e.g., happiness and misery in the world. Or in favor of maximizing something (happiness, aesthetic value, hair length, whatever) only for some subset of humanity, rather than for everyone.


Tim Wilkinson 09.25.10 at 8:26 pm

I agree with Holbo (or with what I think he is opining here). I don’t think these kind of ‘though experiments’ are a source of data about intuitions. The analogy with scientific experimenta crucis is good – and these are precisely not sources of interesting data about the world at large but tests of proposed generalisations. In other words (and other propositions), the fact that these are essentially dialectical, seminar-bound, tools is a feature not a bug: they are a way of trying to make certain aspects of general moral positions both precise and tractable.

For grimmer examples, the fairly recent spate of neocon propaganda about torture and trrrism provides a real world source of inspiration. Of course they are crap as experiments, since far too many uncontrolled variables are allowed (deliberately) to remain uncontrolled, e.g. the distinction between individual action and state policy and the matter of the blameworthiness of of the torturee (cp Nozick on ‘drawing against’ deserved punishment in applying deterrent sanctions). Thay are also obviously (and also intentionally) inadequate in practical terms since they tend to apply to nukes and perfect knowledge but are treated as justifying mass torture of people merely suspected of being vaguely dodgy – i.e they are insufficiently sensitive/entirely beside the point.

These kinds of examples might be quite useful, perhaps as seminar 2 after the basic method is intro’d with trolley stuff, as a way of examining these kinds of defects. Perhaps for example these feed back to the fat man case by way of a more sensitive set of tests for detecting mere squeamishness (how about if the fat man could be placed on the track by pressing a button, or is attached to a pendulum, with a push-button which can be used to stop him when either on or off the track? Further, tweak the no-push default: on, off, unpredictable…

At this point it’s pretty clear that this is not going to work as a intuition pump about brute moral qualities, but as dialectical casuistry, a way of falsifying and/or abducing principles. (If the aim of such experiments were really to find out people’s actual moral attitudes, you would probably need real (or mocked-up) trolleys and people – i.e. you would be doing psychology experiments.)


Matt Austern 09.25.10 at 8:28 pm

In practice, though, are there any consequentialist meta-ethical theories other than utilitarianism that anyone argues for?


ben 09.25.10 at 8:31 pm

“Adiposity Conundrum”, surely.


y81 09.25.10 at 8:45 pm

“In practice, though, are there any consequentialist meta-ethical theories other than utilitarianism that anyone argues for?”

Well, Richard Posner argued explicitly (and many economists, especially on the right of the spectrum, argue implicitly) for wealth maximization, rather than utility maximization. But some would say that this is a very minor wrinkle in basic utilitarianism. (And others, including many here, I fancy, would argue that Richard Posner, and economists on the right generally, aren’t anyone.)


bob mcmanus 09.25.10 at 8:53 pm

2,3:Dostoevsky, William James

Hand them a printout on LeGuin’s “The Ones Who Walk Away from Omelas”

Discuss Rene Girard and scapegoating; Aulen Christus Victor; Iraq & Afghanistan


Tim Wilkinson 09.25.10 at 8:54 pm

Matt Austern – yes, if you have an argument for it, it’s consequentialism (albeit possibly indirect c’ism). This is my opinion and not a standard position though. Also – isn’t it just ‘ethical’ rather than ‘meta-ethical’?

(re me @26: 1. posted without reading whole thread, 2. don’t know why I referred to JH as ‘Holbo’ – maybe just like the sound of it… 3. not terribly cogent, just to preempt anyone else. Am I allowed to do that, without offering an improved version?)


bob mcmanus 09.25.10 at 9:04 pm

Perhaps you could even work from the Trolley Problem to the Difference Principle, and thence to guillotines. Sadly though, the First Principle demands Fat Daddy Warbucks gets his steak (since it appears only violence, throwing him off the bridge, a violation of his rights, will work in real life) as the peasants in the path of the economic express get run over.

But Rawls was a just man.


Paul K. 09.25.10 at 9:11 pm

The truth Tom H.’s claim that you are actively killing the one in the standard Trolley case depends on what it means to “actively kill.” Kamm has argued that while tossing the switch would kill the one–and I think she would concede that the switch-thrower does kill the one–this is a side-effect of saving the five. It is not something that is intended. Since intention is often thought to be relevant in these cases, if actively killing requires an intention to kill, then it’s not clear that the switch-thrower in the standard trolley actively kills.

In my view, any use of the two most standard trolley cases should be accompanied by the loop variant (did Thomson come up with this variant?). This variant is set up exactly as the original standard trolley case, but the track is looped, such that if the one person *were not* trapped on the other track, and the switch was thrown, the trolley would loop around and kill the five anyway. In this case, throwing the switch would save the five only if the one *were* on the other track and only if his body would be sufficient to stop the trolley before it loops back around to kill the five. Many non-consequentialists who are fine throwing the switch in the standard trolley case (because the one is killed as a side-effect of, and not as a means to, saving the five), but who are not OK with pushing the fat man (because the fat man *would* be used as a means), also believe that it is permissible to throw the switch in the loop case. But throwing the switch in the loop case saves the five by using the one as a means. This puts the non-consequentialist (and I am one of them) in quite the pickle, since using someone as a means turned out to be the reason that pushing the fat man was not permissible. I started to read Kamm’s explanation of why it’s permissible to throw the switch in the loop case, but I got confused and stopped.


Timothy Scriven 09.25.10 at 10:08 pm

“In practice, though, are there any consequentialist meta-ethical theories other than utilitarianism that anyone argues for?”

Firstly, utilitarianism isn’t a meta-ethical position. A meta-ethical position regards things such as the meaning of terms like “right”, what makes moral statements true or false ( if anything) so on and so forth. Many have tried to argue for utilitarian theories based on meta-ethics, certainly, but utilitarian theories are fundamentally part of normative ethics. For some reason these things keep being confused on CT, perhaps a clarificatory post JH?

As for the substantive ask, well that’s a difficult question. A vast number of philosophers are consequentalists, but not utilitarians. However, there are just about as many such theories are there are theorists, since every one has their own list of the good. Certain aims that often get chucked in to consequentalist views include respecting rights maximisation, beauty maximisation, friendship maximisation, virtue maximisation, knowledge of the truth maximisation so on and so forth. It wouldn’t suprise me if a majority of consequentalists are non utilitarians.


Randall 09.25.10 at 10:14 pm

just a quick thanks for making the cartoons available—i’ll be using them when i teach the trolley next time! on the switching cartoon, did you intend to make the first of the set of five look like hitler? (at the risk of using weasel words, that would seem to complicate the moral calculus….)


Kenny Easwaran 09.25.10 at 10:45 pm

My worry about the fat man case is that I’m just not as sure that it’ll work to stop the trolley as I am with the switch. And the same is true with the framing one to prevent a riot that kills five. No matter how much you tell me that it’s guaranteed to work, my intuitions don’t believe you and they tell me not to do it. So maybe that’s a reason to ignore the intuitions in these cases?


ben w 09.25.10 at 11:35 pm

Kamm has argued that while tossing the switch would kill the one—and I think she would concede that the switch-thrower does kill the one—this is a side-effect of saving the five.

Any argument that makes this plausible ought also to make it plausible that throwing the fat man into the trolley’s path is a side effect of saving the five as well, or, at least, I can’t see how to distinguish the two.


Timothy Scriven 09.26.10 at 12:33 am

Kenny, any kind of worry like that can usually be answered.

Suppose that the train is not rolling out of control, but is under the complete power of evil von Wicked. Evil von Wicked regularly enjoys running over tied up hostages in this manner, but he has a deep fear of driving his train over fat people (dead or alive), since he had a dream once in which this signfied the end of his evil empire.

You now have two choices 1) Push the fat man onto the track with the five on it, forcing Evil von Wicked to divert to a secondary track with no one on it 2) Do nothing.
Assume that the fall is enough that the fat man will die.

However cheesily you make the point, it is a very real problem, the question of killing N to save P where P>N comes up all the time.


Substance McGravitas 09.26.10 at 12:54 am

However cheesily you make the point, it is a very real problem, the question of killing N to save P where P>N comes up all the time.

Its most depressing form is after the fact rationalization.


dan p 09.26.10 at 1:07 am

Timothy, I think Kenny puts his finger on one reason we *can’t* ignore intuition in these situations. You dealt with his objections by clarifying the hypothetical into an even more ridiculous hypothetical, which makes your last statement unhelpful:
“it is a very real problem, the question of killing N to save P where P>N comes up all the time”

The real problem that people constantly are trying to apply these nonsense hypotheticals to is “killing N persons with very high certainty to save P with much less certainty, where P > N”. I am very loath to kill the fat man because I fundamentally have trouble ignoring my intuition that surely this isn’t the only actual solution. The close relative ticking time bomb situation is similarly absurd in part because of our ignorance about whether torturing will really be the decisive factor in any real case: is he really guilty, will torture really provide the information, is it the only way, etc.

So: bomb a village and certainly kill people who may or may not be fighting against you? Hard problem for reasons that the trolley problems do nothing to elucidate. Our intution is pushing against the unreality of these situations, and expecting them to do much work to clarify decision-making in the real world is foolish. Of more interest would be probabilistic trolley problems:

-the fat man has an 80% chance of stopping the train, and a 99% chance of dying if he succeeds

or ones with competing models:
-there are at least two schools of thought. One says the fat man has a 95% chance of stopping the train (and 99% chance of death if he succeeds). The second says he has a 5% chance of stopping the train and a 99% chance of death if he succeeds. There may be other opinions.

These would get at what is hard in the real world about decision making with uncertainty, which is a major reason our intuition balks at the trolley problems.


Paul K. 09.26.10 at 1:24 am

Ben W.,

I think the difference Kamm has in mind can be seen by noting that in the standard trolley case, you do not need the one person on the spur to be there in order to save the five. In other words, killing the one on the spur is not necessary, as a means, to save the five. In the case of the fat man, you *need* the fat man to drop onto the track in order to save the five. So in the standard trolley, Kamm says that killing the one is a foreseen side-effect, but not intended means. I think she’s right about this.

Are you still not convinced?


John Holbo 09.26.10 at 3:01 am

I agree with Kenny about the ‘where did all this information come from?’ problem – so I guess I disagree with Timothy.

Tell the story this way: ‘imagine that you are a daemon. A daemon is a semi-omniscient, semi-omnipotent spiritual being gifted with perfect knowledge of the probabilities of future events its immediate environment, and gifted as well with the ability to make precise, surgical interventions in that environment with what would be – to an ordinary mortal – an impossibly high degree of certainty. (Daemons cannot do everything, but what they can do, they do perfectly.)’ And NOW you tell the same old stories. I think people’s responses might be a bit different, because they would have an intuitive sense that ‘ethics for daemons’ might be a bit different from ‘ethics for humans’. So the problem with the story as it stands is that it is told in such as way as to suggest we must be daemons, but without actually telling us we are.

An analogy: suppose you ask about whether to shoot a criminal who is about to shoot five innocents, and you tell the story in such a way as to strongly suggest, without saying, that you are a policeman. ‘Your radio will not bring assistance soon enough. You can use your sidearm to shoot the suspect, and you are a good shot …’ You COULD just be some passing civilian who happens to have a radio and gun, but it sounds like you are a cop, so probably the hearer will start to reason like a cop, making it more acceptable to shoot someone ‘in the line of duty’.


John Holbo 09.26.10 at 3:19 am

Sorry for the delays, turning some comments on. (Didn’t notice them languishing in queue.)


Matt 09.26.10 at 3:26 am

29- In one of the few clear, knock-down defeats in a philosophical discussion, Ronald Dworkin destroyed Posner’s argument for wealth maximization in an exchange several years ago, and Posner even admits it. This helped push Posner towards his current anti-philosophy stance, as he still wants to take wealth maximization as the right policy but now claims that he won’t bother to give an argument for it- it’s just a preference he has, or something. It was really an amazing thing to watch.


Paul K. 09.26.10 at 3:49 am

Any citations for that Dworkin/Posner exchange?


geo 09.26.10 at 4:06 am

Aulus Gellius @30: a consequentialist could be in favor … happiness, aesthetic value, hair length, whatever

Timothy @39: respecting rights maximisation, beauty maximisation, friendship maximisation, virtue maximisation, knowledge of the truth maximisation so on and so forth

I’m not sure I see this. Mill (I think it was) once said, in defending utilitarianism, that the point of choosing “happiness” as the thing to maximize was that happiness was the one thing not desired for the sake of something else, whereas everything else was desired because it was assumed to conduce to happiness. (One might want to substitute other words for “happiness”: eg, “welfare,” “well-being,” “satisfaction,” “feeling good,” etc.) This is true of beauty, virtue, knowledge, sex, friendship, health, wealth, etc. One wants them ( in whatever combination) because they conduce to well-being, one’s own or others’. The essence of utilitarianism (and consequentialism, which seems to me the same thing) is that things are desirable, or acts are laudable, to the degree that they make as many people as possible as happy (satisfied, feeling good, imbued with a sense of well-being, etc.) as possible, taking all foreseeable consequences into account, and recognizing that different people are made happy (satisfied, etc.) by different things. In other words, nothing — in particular, natural law, derived from a conception of the human essence, which was the pre-modern criterion of the goodness of acts and things — trumps human desires, wishes, preferences, etc.

In this way of looking at things, there is only one important distinction in ethics: between human and non-human, or naturalist and metaphysical, criteria of goodness and value.


John Holbo 09.26.10 at 4:09 am

“did you intend to make the first of the set of five look like hitler?”

Naw, he was supposed to look more like Apu. Or something. (Now we see the violence inherent in the system of someone who can’t draw, drawing.)


Matt 09.26.10 at 4:33 am

Paul K- it was in the Hofstra Law Review (in a special issue) in 1980 (I believe). Jules Colman’s piece in the same issue was also a wonderful demolition. Posner acknowledged that he had had his clock cleaned in a foot-note in his (awful, I think) _Problematics of Moral and Legal Philosophy_, where he cites Dworkin’s paper, but seems to think we should then just appeal to naked power, or something (he’s not very clear.)


Jack Strocchi 09.26.10 at 4:57 am

Mark Austern @ #32 said:

In practice, though, are there any consequentialist meta-ethical theories other than utilitarianism that anyone argues for?

There were, but they are not mentioned in polite company anymore. Nietzche, Hitler and the Social Darwinists were advocating a form of consequentialist ethics that was not utilitarian. Granted these are not the names that immediately leap to mind for the thoroughly modern ethicist. But they were influential and still haunt our collective imagination.

Survival of the fittest aims at maximizing the power of intelligent predators, whether they be raptors, robber barons or SS stormtroopers. So it is definitely a teleological consequentialist philosophy.

But its not utilitarian. Romantic nationalists contemptuously dismissed utilitarianism as a shopkeepers mentality unfit for higher breeds.The predators were not going to have a happy contented life, what with the ceaseless struggle to survive and dominate.

Its not always clear that we have said Goodbye to All That. The ethic driving the stockmarket – profit maximization – looks utilitarian given its basis in neo-classical economics. But this appearance is deceiving.

It always assumes that there is a person behind the profession. Once upon a time there was an individual there, now its institutional. The notion of a company having feelings is a bit of a stretch.

One day “the agent” could be purely instrumental ie a computer. You would have a hard time selling the greatest happiness of the greatest number of microchips as “utilitarian”.


Jack Strocchi 09.26.10 at 5:50 am

John Holbo said:

the problem is not that Anglo-American, analytic-style philosophers (call them what you like) are severely autistic but that they are mildly whimsical. This style of example-mongering complements a characteristic style of seminar room cleverness. There’s an intellectual ethos that goes with this social ethos; and, if you ask me, there’s a lot to be said for it.

I am not sure that the analytic movements obsession with fanciful thought experiments has done much to improve the standing of philosophical ethics. It makes philosophy look frivolous, not helpful given the utility of any kind of philosophy is not always obvious to the long-suffering tax-payer.

I can see the sense of Rawls “veil of ignorance”, and even Nozick’s “experience machine”, as a way of testing our moral theories of Right against intuitive notions of Good.

But the “trolley car” and “fat-man” thought experiments are more plausible as Goon show scripts rather than moral experimental design. Hard cases make bad laws.

FWIW, I am a re-constructed evolutionary utilitarian (Hayek’s version) . Naturalistic fallacies aside, you have to have some way of measuring progressive performance. But performance is always relative to the game and in evolution the goal posts are always changing.

Utilitarianism at least makes use of double-entry accounting, At the individual level we all have the natural capacity for pleasure and pain which only a fanatical deontologist could deny. And at the institutional level its impossible to run any organization without taking account of costs and benefits.

Granted one can always poke holes in utilitarianism’s tendency to aggregate which may sacrifice persons. But each is to count for one and none for more than one is a pretty good start.

It seems obvious to me that in evolution, the notion of the Good must precede moral conception of the Right. Much as we evolved the sense of tastyness in order to motivate the hunt for high energy foods. But menus and recipes change, depending on circumstance. So one should be an Absolutist about the Good and a relativist about the Right.

“Feel-good” is natures way of getting the individual do more of what gets him ahead. “Do the right thing” is cultures way of institutionalising the rules that promote the good

Ideally one should “Feel Good” by “Doing Right” which “Gets You Ahead”, at least in the perfect state envisioned by original utilitarians. Thus do “natures gentlemen” eventually get their Knighthood.

But this scenario requires a society of saints. Which is not exactly the state that secular liberal utilitarianism is leading us towards.


Timothy Scriven 09.26.10 at 6:50 am

The problem with complaining about thought experiments being “frivolous” without backing it up with some theory as to how this could be distorting what is found in a pernicious way is that it makes philosophy into a kind of Belles Lettres. Other than a preconception that it is somehow very wrong to go about it in such & such a way, I just don’t see what the problem is. Would examples that reflected the scope & grandeur of the questions somehow lead to better or clearer results? If you think so, fine, but that has to be argued for , and once you start to play that game you might soon fine you’ve started to play ours.


John Holbo 09.26.10 at 7:19 am

Timothy, what do you think of my daemon objection?

In general, your objection to making philosophy into belles lettres seems to me question-begging: the concern is that whimsical examples are, already, a minor literary genre in their own right (hence this style of philosophy is, in a sense, already belles lettres, in a small way). And that this may affect our response. This may be right as an objection or it may be wrong, but it is no sort of objection to this concern that it would turn philosophy into belles lettres. This just restates the concern itself, no?

I’m not sure that scope and grandeur are what we need. It might be that we need more realism. It might be that the mild absurdity of these environments keeps us from applying ‘thicker’ conceptual frameworks that we might bring to bear on realistic cases. The set-up is so bizarre that we don’t really have any notion of social expectations or standard social roles. (Is this a world in which it is common for people to find themselves suddenly so knowledgeable about this sort of stuff? If not, then why do I know it, in this world?) Weird cases may beg the question against certain forms of virtue ethics, by screwing up the background in ways that foil our attempt to conjure up a sense of what ‘good judgment and wise action’ would be like, under the circumstances.


bad Jim 09.26.10 at 8:10 am

Just how distinct are the ticking time bomb and trolley dilemmas once you stipulate that torture is effective? If you need to vary the degrees of moral discomfort in the former instance, you have threats ranging from pain to impairment to disfigurement to death, and also the death of loved ones, a wider range of choices than the latter selection of an anonymous victim, a chosen victim, and a modest assortment of innocents.

When I look at the fat man on the overpass, I think (1) the fall will kill him, even if he misses the tracks, (2) if he hits, he might kill one or two on impact, (3) if he has sufficient mass to stop the train, he’s probably only saving the last two or three on the tracks, and (4) the trolley’s engineer and many of the passengers are at least likely to be injured in the impact.

Most recent arguments about torture actually concern torture; as far as I know, the U.S. hasn’t recently researched fat man deadfalls for trolley traffic control. Apart from practical applications, though, the scenarios appear precisely parallel.


Bruce Baugh 09.26.10 at 8:38 am

I’ve been thinking about the matter of whimsy. This feels significant to me.

For someone coming out of my kind of bourgeois environment, direct power of life and death is fairly rare. Death comes at a distance (soldiers on assignment somewhere else), or as the result of untreatable calamity (illness, t o), and that’s about it. Any deliberate and immediate killing is very unusual, and very serious, and taken to be a likely-traumatic matter, and like that. Being goofy about it feels like a really serious mis-match of subject and tone.

Now there’s a lot you could say about responsibility for individual and mass death not recognized or not accepted, and I’d agree with a lot of it. Death and misery for a lot of others does indeed go with bourgeois life. But the issues there don’t seem to me to correspond at all to these thought experiments fusing comical flights of fancy with what’s always struck me as Book of Questions-like tough-guy implications.

I do admit that it’s really hard to construct pedagogical examples that are both clear and worth bothering with – in fact I don’t think the talent for that has any particular correlation with, say, the talent to come up with philosophical insights of general usefulness.


Bruce Baugh 09.26.10 at 8:43 am

Shorter me: For me and a lot of folks I know, the implicit cultural norm is that funny deaths are the ones happening to people not like us and/or far away from us. To the extent a postulated death feels goofy, it’s going to tap into reactions about those kinds of deaths rather than ones closer and more significant-seeming. This probably skews ensuing judgments.


tomslee 09.26.10 at 10:48 am

Why is the problem never described from the point of view of the fat man. If you were he, would you jump? Would you shout “quick, push me over?” The regular phrasing falls short of having the courage of its convictions – under the guise of making it real, it keeps the protagonist at a nice distance from being really involved in the problem.


Alex 09.26.10 at 11:55 am

@61: nice.

More broadly, this comes to mind: Conservatism – we shouldn’t intervene in the ancient tradition, wrapped around the efficiencies of the market, that runaway trolleybuses crash into crowds. Communism – sure, push the fat guy under the trolleybus. got any more? Socialism – why are these trolley buses running away down the hills? todo: new inspection standard for trolley brakes.


Henri Vieuxtemps 09.26.10 at 1:05 pm

@61, you’re missing the point: most people’s intuition already says that pushing the fat man is wrong; even without emotional phrasing.


Paul K. 09.26.10 at 1:40 pm

Tom Slee, Critics of consequentialism often point to two features of the view that seem counterintuitive: first, it seems to require us to violate intuitively compelling moral rules (e.g. rules against pushing people into trains); second, it seems to place unreasonable demands on our personal time, energy, attachments and, yes, sometimes lives.

One reason why the standard Trolley problems involve pushing and not jumping, I suspect, is that there is no moral stricture against sacrificing oneself to save others. So when you jump, the central issue it raises was whether you were, in fact, obligated to jump. That’s a real question worth addressing. But it leaves out at least one element of the debate that some people will want to focus on (namely the violation of commonly recognized moral strictures).


Paul K. 09.26.10 at 1:40 pm

P.S. Thanks for your wonderful book!


Witt 09.26.10 at 3:01 pm

People aren’t reporting their ‘intuitions’. Period. Rather, they are telling you what they would say, once you’ve set the mood for a good, seminar-style go-round. The examples are supposed to take you out of the seminar room, imaginatively. But plausibly they do the opposite – are really designed to do the opposite. They export the socially and intellectually distinctive seminar room atmosphere to a whole, slightly off-kilter world.

Between this and your last paragraph, I’m still not convinced that trolley problems are a desirable frame for discussion, but I’m fully convinced that teaching and discussing them JH-style would be a marked improvement.

LizardBreath at 28:
If you’re asking students for their intuitions, you’re asking them to make a general assessment of everything they know and feel about the situation, and expecting them to be able to discard their sense of the probable and accept both that the fat man’s body would stop a trolley that would be unimpeded enough by four (presumably leaner) bodies that it would still kill the fifth person, and that the student would be certain enough of this to act immediately on that basis. Even though that’s the stated premise of the question, it seems implausible enough to me that it’s going to throw off answers.

To follow up on the bolded portion, another problem for me is that my assessment of the situation doesn’t stop with the moment of impact. What’s the difference for the trolley driver and passengers if the collision is caused by hitting one person rather than hitting five? Are they going to be injured at the same rate and in the same way? Do I think they will feel worse, going about their lives afterwards, knowing that part of the “accident” ws intentional, rather than purely mechanical or happenstance?

If someone [me] pushes the big man, is s/he going to spend the next five years of life with a massive liability lawsuit hanging over their heads, as the insurance companies battle it out?

Obviously part of the reason for setting up thought exercises is that you’re supposed to abstract all of these problems away. But when I’m asked for my “intuition” (really, more like my reasoned reaction) to a situation like the above, then my pragmatic sense of reality comes to the fore.

Yet another wrinkle is that hypotheticals like these require: 1) that the student believe the professor is presenting them in good faith, and 2) that discussing them can be done in a sealed seminar room, with no leaky consequences for the real world.

It’s not hard to imagine someone committing #1, and #2 immediately brings to mind “should killing a fetus in the act of homicide to the mother be a separate charge of murder” — a question which could be framed as a safe classroom hypothetical, but in fact is politically used as fodder for laws that stake out ground for a claim of fetal personhood, and consequently a climate of decreased reproductive rights for women.

Shorter me: Doing trolley problems requires a lot of trust on the part of the students, in ways that the traditional framing of the problems often seems designed to elide.


Witt 09.26.10 at 3:04 pm

Whoops, my phrasing was clumsy above. I meant “It’s not hard to imagine someone committing #1″ is meant to convey “It’s not hard to imagine a professor presenting a problem in bad faith.”


bianca steele 09.26.10 at 3:40 pm

LizardBreath wrote that, for some students, the way the trolley problem and other thought experiments is introduces seems to be “expecting them to be able to discard their sense of the probable” and to proceed on trust that the problem has to be couched in the way it is.

I think this is key. In some situations, a teacher absolutely wants students to question their own sense of the probable. Howard Gardner has used the example of students’ intuitive sense that the physical world follows Aristotelian physics, and how teachers of physics have to inculcate in them a new sense that Aristotelian physics is wrong, and have them replace it with Newtonian physics. Carol Gilligan’s theories can be used in support of the thesis that women are innately less Kantian than men (that’s a very rough account of something Seyla Benhabib has written). But, I think, the issues are not exactly the same. For one thing, I think students who are very good at abstract symbol manipulation in physics, math, logic, and computer science are often more likely to find objections to philosophy thought experiments that involve real-world probability. I think it’s probable John Holbo is right that an “ethos” has something to do with it.


Tim O'Keefe 09.26.10 at 3:50 pm


One non-utilitarian consequentialist theory is ethical egoistic hedonism, according to which you should maximize utility (balance of pleasure minus pain), just as with utilitarianism, but not the utility of all people weighed equally, but simply your own utility. So the shape of a particular consequentialist theory depends not just on which consequences you seek to maximize (or minimize), but the consequences for whom. (Within utilitarianism, you can query whether we should count equally the pleasure and pain of all humans, or of all sentient creatures, human and non-human. Nowadays I believe the latter is the more popular option among utilitarians.)

I might as well also mention (as I believe it hasn’t been mentioned before) that one could take the balance of desire satisfaction (minus desire frustration) as what we should seek to maximize. However, folks who put forward the agent-relative version of this (i.e., the version according to which it’s your desires that count) usually put it forward as a theory of rationality, not of ethics.


bianca steele 09.26.10 at 3:51 pm

There’s a missing transition in that comment, but I’m not sure what it should be. Sorry.


geo 09.26.10 at 3:57 pm

Timothy @56: philosophy … a kind of Belles Lettres

Hasn’t Rorty already established this?


John Holbo 09.26.10 at 4:05 pm

“Hasn’t Rorty already established this?” Well, he said it, anyway. I don’t mean it in quite the same sense that Rorty said it.


SusanC 09.26.10 at 4:18 pm

There’s a genre of psychology experiment where the unfortunate experimental subject is put into a situation they wouldn’t normally encounter. (e.g Milgram’s experiments on obedience: the typical experimental subject for this — from a western democracy — doesn’t normally have to deal with someone official-looking asking them to electrocute a prisoner).[*]

The subjects’ behavior in these contrived sitations can tell us something about the cognitive shortcuts people are using–which is fine.

More controversially, there’s often an implication that people’s heurisics are wrong or bad, because they gave the “wrong” response in the experimental situation. I don’t entirely agree with this–to be effective, our heuristics only need to be right most of the time, in the situations we normally encounter. Failure under wierd conditions isn’t a practical problem.

Philosophical thought experiments are even worse.

1. What someone says they would do when asked in a seminar room could be very different from what they would really do if you put them in a (simulation of) that situation. (Though your department’s Institutional Review Board might have an issue with tricking a subject into thinking they were faced with the trolley problem).

2. They also have the problem (shared with some psychology experiments) that the situation is unusual. Does it matter if our heuristics break down in unlikely situations?

I’d extent this object to some of analytical philosophy’s thought experiments about language: it’s not a problem if our language breaks down entirely in unlikely situations (e.g. is unable to describe the situation, or cannot describe it unambiguously). Language only needs to “work” (enable effective communication) in typical situations. It can do this even if people don’t know/don’t agree on what utterances “mean” in unlikely circumstances.

[*] Yes, I know, Milgram’s experiments were motivated by Nazi Germany, which was in a western “democracy” not that long ago. So there’s an argument that could be had about whether Milgram’s experimental situation is actually unlikely. But, probably, the experimental subject has never encountered it before. (And in other experiments, the situation is definitely unlikely).


SusanC 09.26.10 at 4:22 pm

P.S. I know a lot of people who are into BDSM, which makes me a little skeptical about the implied morality of the experimental subject in either the Trolley Problem or Milgram’s experiments (go on, you’re a sadist: you know you’ld *enjoy* shoving the fat guy under the trolley/electrocuting the prisoner)


ben w 09.26.10 at 5:36 pm

In the case of the fat man, you need the fat man to drop onto the track in order to save the five. So in the standard trolley, Kamm says that killing the one is a foreseen side-effect, but not intended means. I think she’s right about this.

The means of stopping the train isn’t killing the fat man, the means is putting something really heavy in front of the trolley. The only heavy thing to hand happens to be a very fat man and I foresee that he will die when the trolley hits him, and that is too bad, but that he should die isn’t part of the means any more than the death of the man on the other track, which is also too bad, is part of the means when I flip the switch.


Russell Arben Fox 09.26.10 at 5:55 pm

Not having the time to read all the comments, I can only submit my experience as a support for what John says in the original post; I decided to use Sandel’s book (in the place of Harry’s, fwiw) in my Christianity and Social Justice course this fall, and we started right out with the Trolley Car, supplemented with clips from Sandel’s lectures. The discussion went very, very well. I think I’m probably more suspicious of the intellectual and social ethic that is presumed by this kind of philosophical reasoning than John is, but the truth of the matter is that I’ve rarely had a introduction to utilitarianism lecture go quite so well before. The Trolley Car works.


roac 09.26.10 at 6:09 pm

Where I am in relation to the switch? Do I have a direct view of the switch and the track? How do I know for a fact that the people tied to the track are people and not lifelike dummies (maybe the bad guys are trying to fool me into killing the one live person at risk. Did I watch while the bad guys tied them all up? If so, why was I totally unable to do anything to prevent them? Or at least to call up Trolley Central and get the system shut down? Did the bad guys lock me in the control tower first, cut the phone wires, and tie me to a chair so that all I can move is the finger that is hovering over the switch? Why did they go to all that trouble instead of just killing me? It seems likely that this is an elaboration of the Milgram experiment, and that I, and not the people on the tracks, am the object of the whole operation. So I am following my intuition, which tells me to leave the damn switch alone.


tomslee 09.26.10 at 7:11 pm

I find LizardBreath @28 persuasive, coupled with the fact that someone of my upper body strength is likely to push without effect, leaving me rather embarrassed in front of a big and angry man.

Assuming that you do accept all the assumptions of the problem, my son has sat through this in class, and is frustrated that people don’t see SusanC @73′s point (“our heuristics only need to be right most of the time, in the situations we normally encounter. Failure under weird conditions isn’t a practical problem.”) As he puts it, “So your intuition is wrong. Get over it. End of discussion.” And the story of modern physics, of course, is the realization that one intuition after another from our slow-speed, moderately-sized every day experience does not apply in conditions other than those under for which it was developed and must be jettisoned.

[And Henri@63, Paul @64: agreed. Thanks for clarifying.]


johnm 09.26.10 at 7:27 pm

This is probably just useless musing on my part, but I wonder how the experiment would differ if it weren’t faceless or anonymous people tied to the tracks.

For a trite example, suppose it’s 1942 and Hitler is one of the 5 the trolley is barreling toward (the other 4 are still anonymous). And for some reason Hitler will get away if the trolley doesn’t get him, I guess.

Or suppose Mother Theresa or some other universally recognized as virtuous person is the one, and the trolley is headed at the one instead of the 5. The 5 are all bad people (say, bin Laden, Hitler, Stalin, Pol Pot, and Lindsay Lohan).

You get the idea. Variants of good/evil/important people on the tracks.


John Q 09.26.10 at 7:29 pm

Missing from the trolley problem is the nature of the individuals tied down.

If there are five Al Qaida or Aryan Brotherhood members in the path of the trolley, and one university professor or airline pilot tied to the alternate path – yes, I would throw the switch.

It gets a little murkier if there are five woman Al Qaida members who each have three very young children dependent on them against one childless airline pilot.

The utilitarian approach assumes an equal value for each individual in the problem, which just isn’t the case in real life.


bianca steele 09.26.10 at 7:50 pm

Maybe I’m being whimsical, but if the point of these thought experiments is taken to involve getting people thinking like a philosopher (as the OP says), and it also involves consequences (it isn’t that throwing a switch is considered always bad but only because it kills people, and it isn’t that pushing a fat man off a bridge is considered always good but only because it saves people), it seems like that could be confusing. If thinking like a philosopher, you want to make your categories as broad as possible. So, as suggested, why not generalize throwing a fat man off a bridge to causing a man to fall from a bridge onto trolley tracks whenever there is a trolley on path to run people over . . . and then why not generalize further, to throwing yourself into the path of any trolley you see?


M31 09.26.10 at 10:02 pm

I like to pose this one to folks who insists that ‘life begins at conception’, i.e., a fertilized egg = a human being:

You are in a fertility clinic, and it’s on fire!!!!! You can only grab one thing to save, which do you choose:

1) a petri dish containing 5 frozen fertilized eggs, or

2) one live 2-year-old child

I keep waiting for this question to be asked of some high-profile right-to-lifers, but somehow I haven’t heard it yet.


qb 09.26.10 at 10:21 pm

SusanC @ 73: What someone says they would do when asked in a seminar room could be very different from what they would really do if you put them in a (simulation of) that situation.

So what? Thought experiments are supposed to help us identify intuitions about what we think we should do in various situations. No one denies that what we would do and what we should do can often come apart, especially in situations where doing what we believe to be the right thing involves acting against our natural or socialized inclinations.

Does it matter if our heuristics break down in unlikely situations?

It does if there is a better heuristic available, i.e., one that doesn’t break down in as many unlikely situations. One of the reasons to resort to thought experiments in the first place is that the dominant ethical theories generally agree about what we ought to do in ordinary situations; comparing their responses to less ordinary situations is a way to tease apart their different implications.


Sarang 09.26.10 at 11:22 pm

Hm… green eggs and ham.

Would you choose to pull the lever?
No I do not think I’d ever!
I would not choose to pull the lever.

Would you push the fat chick over?
I would not push the fat chick over,
I’d certainly refuse to shove her.
I would not choose to pull the lever,
No I do not think I’d ever.

Would you pull it for the orphans?
How about endangered dolphins?
No orphans. No dolphins. Etc.

Would you save the serial killers
Or the authors of bad thrillers?
No killers. No thrillers. Etc.

What if they’d have cured your cancer?
Would that, maybe, change your answer?
No cancer. Same answer. Etc.

What if they were Mao and Stalin
Or autistic infants bawling?
No Stalin. No bawlin’. Etc.

At this point I’m sort of running into the limits of my philosophical education so I’ll stop here.


Natilo Paennim 09.27.10 at 12:15 am

Aha! Now the poetry has begun:

I set out one morn, with nary a care.
Streetcarwards I went, oh! but the folly!
I looked down the tracks, and I got a scare:
People were bound, in front of the trolley!

Five tied to the tracks! Goodness, oh golly!
Who’d play such a prank, these folks to restrain?
Then I spied a philosopher, jolly:
“He’s doomed them,” I thought, “He must be insane!”

“Here’s your choice,” he said, “what’s most humane:
“You may flip this switch, and change up the track,
“To one where I’ve tied a man I disdain!
“One life for five, and a horrible smack!”

Now I will advance this question to you:
With no clear moral choice, What did I do?


Paul K. 09.27.10 at 12:59 am

@Ben W. (#75): you’re right: you don’t need the fat man to die. But you do need him to be there, and you do need to place his life at risk in order to save the five. This is a key difference from the Standard Trolley, as you simply do not need the one to be on the other spur in order to save the five. So the need to use the one as an intended means (which is really what most people find wrong about the Bridge Variant) is there in the Bridge Variant but not in Standard Trolley. In Standard Trolley, you do not use the one on the second spur as a means. You don’t *use* him at all. This is why it might be misleading to characterize his death as an “active killing,” as Tom H. did.


Daniel Thiel 09.27.10 at 1:13 am

Hi John!
Lovely pictures. You were my t.a. at Berkeley. I’ve always admired the fact that in the midst of your dissertation, you audited a Shakespeare class and came out to see my band play. Any way…trolley problems… Have you read Steven Stich’s work on these? It is very interesting…experimental philosophy…subjecting “universal” intuitions to empirical verification. Worth checking out. Hope all is well. Did you publish the ABCderium?


john c. halasz 09.27.10 at 3:55 am

Geo @ 51:

I find your account there thoroughly confused. It’s a commonplace that moral acts involve a possibly painful self-discipline , such that the “right” (or “just”) acts come at the expense of personal happiness, whereas personal happiness often inclines one to “wrongful” (or ‘unjust” or “selfish”) acts. That is much of what moral controversy is about. I wouldn’t want to swallow the Kantian hog whole here, but that goes to why he stigmatizes embodied. empirical motives as “pathological”. Of course, Kant is under peculiar compulsions due to the way he sets up his whole problematic. Kant is a “phenomenal” determinist and thus must likewise be, in part, a moral skeptic. Hence the only sign of any actual “freedom”, (identified dubiously with “autonomy”, itself identified dubiously with the moral law), is the “Ehrfurcht”, (reverence or awe, combining literally honor and fear), for the moral law, and thus the painful pleasure that on experiences in conforming oneself to the moral law, (which one has voluntarily at once legislated and imposed on oneself). I wouldn’t want to identify myself with the purist/rigoristic/formalistic nature of the Kantian account, but it does capture something of the phenomenology involved. Likewise, though the internalism/intentionalism of Kant’s moral account, (its excess of Lutheran inwardness), is hard to countenance, (and yields an oddly abstracted, disembodied account of human agency, by implication), his notion that intentions rather than consequences are the locus of moral responsibility and culpability, does capture part of the “phenomenology” or the intuitions that most here might share. Hence too his claim that the “good will” is the only “thing” that is good in itself and, as the only “true” end-in-itself, underwrites moral “disinterestedness”. Most of us, I’d guess, would reject such “pure” intentionalism, since acts have conditions and consequences, as well as, intentional sources, and it would be difficult to specify intentions without resort to their conditions and consequences. But it doesn’t at all follow that “consequentialism” (= utilitarianism) would be the only alternative conception or adequate to our various intuitions on such matters.

But what I do want to defend from Kant, leaving aside much of the rest of the apparatus of his thinking, is his basic intuition or recognition of the de-ontic status of norms: i.e. that norms can not be “grounded” or “justified” either on the basis of some pre-given metaphysical-teleological order, nor on the on the basis of some set of naturalistic facts. (Since norms are counterfactuals, they are evidenced as much by their violations as their observances or enactments). That is a far more “radical” account of the problematic of “morality” than what you adduce from Mill, (with its dim echo of Aristotle, though, as many have pointed out “happiness” is very much an 18th century empiricist notion, which is not the same as “eudaimonia”, with “human flourishing” often suggested as the translation, though “well-being” would be a serviceable current word, if well-faring might be too old-fashioned and contaminated by modern meanings, though I’d prefer the literalistic “well-spiritedness” to capture the “original” intent). At least, I think that enters more “deeply” into problematic modernity, wherein a formal-rational systematic prescriptive ethics breaks down and becomes “no longer” possible, which is the burden of much of the crisis mentality of 20th century philosophy, for better or worse. Maintaining otherwise tends toward academic pettifoggery.

As to utilitarianism, (which again is not the some as “consequentialism”, whatever that might means, since consequences, foreseen or not, can’t simply be ruled out of the subject-matter), it is subject to a number of obvious objections. It tends to reduce ethical issues to a form of pseudo-economic reasoning, in which all goods, needs, desires, ends, are assumed to involve a single homogeneous criterion, “utility”, which itself involves a simplistic, binary account of pleasure and pain, carrots and sticks, as the root of motives and ends, desires, ultimate or proximate, and thus emphasizes the efficiency, optimization and accumulation of means over any validation of ends, hence confusing deliberation with calculation, etc. But the most basic objection is that it fails to provide any adequate criterion of recognizing the separateness of persons. If ethics concerns acts and material or symbolic exchanges with respect to relations with others qua other, i.e. irreducibly separate selves/agents, the the whole terrain of conflicts between agents and between potentially incommensurable “values” and ends. and their potential resolution, which motivates ethical inquiry, in the first place, is evaded or elided. “Happiness” just amounts to a name for assuming the consequent, avoiding awareness of the “tragic” potential involved and the need to differentiate the sort of “higher” desires involved in the pursuit of “right” or “justice”. And it also misses the fundamental ambivalence of human desire, whereby goods and bads are not simple binary opposites, but often all too proximate, such that good and bad can exchange places in processes of normative transformation in recognizing “the good”.

As to the actual, general topic here, alas, once again, the “trolley problem”, I belong to the statistical majority in intuiting the “switch” case permissible and the “fat man” case not. But I think the basis of the intuition can be articulated: “fat man” involves not an active deliberate act, which “switch” does not, but rather a direct, immediate relation (to both the act and the other), whereas “switch” is at once more distantiated and mediated by an apparatus. (In fact, I thought fMRI atudies had been done on the problem, indicating, unsurprisingly, that differently located brain areas were involved in the two case, for all that I dislike facile extrapolations from mere brain images). But if Analytic sharpies insist on providing examples as “intuition pumps”, shouldn’t they maybe be engaged in multiplying different examples, (since examples in philosophical reasoning are always tendentious, whether consciously or not, as both Wittgenstein and Adorno, in their different ways, took pains to point out). Umm… that might at least move the discussion forward, rather than merely illustrating received wisdom.


ben w 09.27.10 at 4:02 am

So the need to use the one as an intended means

Oh, well, I wouldn’t disagree with this characterization: you do use the man as a means. But killing him isn’t the intended means, which is what you said. It’s not as if you’d say “it won’t work then” if you found out that he was not going to die.


Tim 09.27.10 at 4:06 am

I always thought the point of the trolley problem is that there are no right answers. Because there is no right answer, you can always ramp up the absurdity until strict utilitarian answers are met with disgust (killing someone to harvest their organs and save five people) and the saps who won’t throw the switch are at a loss to explain why 5 people dying is better than one.

The trolley problem proves that the social context, i.e. how you will be perceived by others afterwords, is more important than actual outcomes or motives and that objective ethical systems are illusions. Hence, we have 86 hair-splitting comments and no satisfying answers.

BTW John, I love the cartoons. They illustrate the absurdity of the problem brilliantly.


Salient 09.27.10 at 4:07 am

I have nearly-complete confidence in the fact that flipping the track switch will have the intended effect. I have no confidence whatsoever that the fat guy will stop the train.

And — the important part — I don’t care how many times I get told that I know the fat man will stop the train. I can be told to assume this and intuit accordingly. Doesn’t matter. It doesn’t even matter whether or not the word know is printed in italics, for my edification. Might as well tell me to physically defy gravity, and then intuit accordingly. Go ahead and tell me I’m supposed to pretend I have complete confidence. I cannot pretend this. I can “pretend to pretend” and go along with the discussion. But I cannot convince myself that pushing a fat man onto a train track will reliably stop a train, or even have more than a slim outside chance of stopping a train.

So if anything, for me, the trolley problem illustrates the limits of my own imagination. I can’t find the implausible sufficiently plausible to exercise intuition on it. What it specifically tells me is that, if I were completely confident that pushing the fat man would stop the train, then I would have to be the kind of person who could conceivably be completely confident about that sort of thing, which would be a very different person than who I am: so very different that I can’t even attempt to see through their eyes or adopt their perspective.

Yet I am completely confident that the hypothetical track switch will work, and can intuit accordingly. So it’s not a question of “complete confidence” or near-complete confidence having no meaning to me. It’s a question of what my brain is capable of entertaining.


Timothy Scriven 09.27.10 at 4:26 am

If the point is that, in the real world we can never be sure that the dillema is unavoidable, sure, that’s true- we can never be SURE the dillema is unavoidable. However, that is not necessairly morally relevant. If I were Superman or MacGuyver I might be able to to save everyone somehow, however I’m not, so I’m just going to have to kill the fat man and live with the consquences, or do otherwise and live with the consequences..

Real world analogies of this come up all the time, and sometimes the agents in question have no escape. Generals commanding solidiers to hold the rearguard. Beauracrats juggling various medical budgets. I think I’ve even heard of a real world, honest to God railway problem just like this somewhere. The fact that a paticular moral dillema isn’t specified finely enough to make it clear that there is no way out misses the point.

Ultimately, when we approach a moral choice we have to frame it against a background of assumptions. We are not certain of these assumptions, but if we’re hoping to avoid all 5 v 1 dillemas, or equally awful Sophie’s choice style problems, simply stating that there might, in principle, be another way won’t save us.

John seems to be worrying about the friviolity of the examples in and of themselves. I don’t see the intution. Should we frame them in gravity & grandeur to match the awesome life and death nature of the questions? What would that acheive? Is it just about the taxpayer’s dollars? If so, is this really just a kind of propoganda exercise? I think there are grounds for believing that my work, in the long run, is worth the money that funds it, but it’s not my job to convince the taxpayer about that.


Salient 09.27.10 at 4:29 am

Suppose there’s an unconscious person stuck under the base of the trolley switch (no need for them to be fat) and pulling the switch down seems very likely to maim/kill them (e.g. by breaking their neck).

Would anyone hesitate to run over and pull the switch, even though it’s clearly going to result in gore and death, with consequences even more in-your-face than the guy-pushed-onto-tracks example?

(Here, the implausibility is shifted onto the likelihood of death for the fat man, instead of onto the reliability of the fat man as a train-stopping device. Maybe I’m wrong in supposing nearly everyone would go pull the switch in this case, but if I am wrong, then I would think this hypothetical is superior to the classic example because it would better illustrate that the problem’s not with the limits of what is conceivably plausible to us. But I suspect I’m right and that people would pull the switch, even though it’s basically set up to be a track-switching guillotine.

Also, there’s probably a lot of student benefit to be gained from having students attempt to construct revised hypotheticals like this, tailored to address specific given objections, but that probably gets pointed out all the time.)


John Holbo 09.27.10 at 4:49 am

“Should we frame them in gravity & grandeur to match the awesome life and death nature of the questions? What would that acheive?”

Timothy, I’ve responded to this upstream – #47 – and am curious what you think of my response. I think gravity and grandeur might not be it, but flatter realism might be. (Grandeur isn’t the only alternative to mild whimsy, after all.)


Sebastian 09.27.10 at 5:05 am

So what is the moral valence of getting the switch problem wrong and accidentally switching to the 5 person track, thinking you were switching to the one person track?

I suspect most people will excuse the switcher on intentional grounds, but it should reveal something about certainty and belief that might be interesting.


Timothy Scriven 09.27.10 at 5:15 am

John- I totally agree that it’s worthwhile trying out thicker cases which capture the nitty gritties of day to day reality to see if different results come about. Most of my word is in experimental philosophy at the moment, so I find the idea of using emotionally and socially richer examples quite exciting really.

My concern about turning philosophy into Belles letters was more about other commentors like Storrochi than with what you were saying. There is a sentiment that I get from my friends elsewhere in the art department that philosophy just isn’t literary enough.


zamfir 09.27.10 at 6:20 am

Until these pictures I had never imagined people on the trolley, and definitely not a driver.


Steven desJardins 09.27.10 at 8:01 am

There’s a utilitarian objection to pushing the fat man, which is that principles have utilitarian value. If fat men know that they’re subject to summary execution for the common good when walking over bridges, they’ll be less likely to walk over bridges; every time a fat man chooses to avoid a bridge, he loses something of value to him. This consequence seems silly—how many people would consider this possibility when deciding whether or not to use a bridge?—but that’s a consequence of the silliness of the scenario and a reason not to use silly scenarios: the principle “fat man should not be subject to summary execution when crossing bridges” is a specific application of a far broader principle of justice, and if society is willing to sacrifice people in this case, it implies that society would be willing to sacrifice people in other cases. I’d rather not live in a world where I have to arrange my life so that my death won’t be convenient to others.


Alex 09.27.10 at 8:40 am

@79: because philosophy always tastes better with Hitler!


Henri Vieuxtemps 09.27.10 at 8:41 am

Salient, to address your objection they invented another story: fat man blocking the exit from a cave (number 4 here). And you’re right, 74% feel like blowing the fat guy into pieces.


Henri Vieuxtemps 09.27.10 at 8:45 am

…although maybe it’s because in this story you’re one of the potential victims, trapped inside the cave?


Earnest O'Nest 09.27.10 at 9:50 am

97-98: might become real soon, in Chile.


ajay 09.27.10 at 10:17 am

99 is actually true, given the emphasis that’s being put on the miners losing weight in order to fit up the rescue shaft.


Ginger Yellow 09.27.10 at 10:33 am

The trolley problem baffles me as I have the apparently non-standard response that you should kill the one to save the many in both situations. I really don’t see the difference between the two, and you are definitely actively killing the one in both cases. Now I have seen broadly analagous experiments constructed where my intuition flips to the norm, but to me all that really does is highlight the problems with thought experiments Kenny elucidates – an unrealistically high degree of certainty about the consequences of extremely unusual actions. In any real world equivalent I’d be much less inclined to push the fat man because I’d have much less certainty that it would save the five. And I suspect, with Kenny, that this is what really underlies the “normal” response to the trolley problem.


ajay 09.27.10 at 10:41 am

Real world analogies of this come up all the time, and sometimes the agents in question have no escape. Generals commanding soldiers to hold the rearguard.

An issue addressed, in fact, by the renowned philosopher, hack writer and glovemaker W. Shakespeare.


But if the cause be not good, the King himself hath a heavy
reckoning to make, when all those legs and arms and heads, chopp’d
off in a battle, shall join together at the latter day and cry all,
“We died at such a place”; some swearing, some crying for a
surgeon, some upon their wives left poor behind them, some upon the
debts they owe, some upon their children rawly left. I am afeard
there are few die well that die in a battle; for how can they
charitably dispose of anything, when blood is their argument?
Now, if these men do not die well, it will be a black matter
for the King that led them to it; who to disobey were against
all proportion of subjection.


So, if a son that is by his father sent about merchandise do
sinfully miscarry upon the sea, the imputation of his wickedness,
by your rule, should be imposed upon his father that sent him; or
if a servant, under his master’s command transporting a sum of
money, be assailed by robbers and die in many irreconcil’d
iniquities, you may call the business of the master the author of
the servant’s damnation. But this is not so. The King is not
bound to answer the particular endings of his soldiers, the father
of his son, nor the master of his servant; for they purpose not
their death, when they purpose their services.


Stuart 09.27.10 at 3:08 pm

Personally I have always seen the “fat man” example as inane – it is like a maths teacher asking everyone what does 1+1 equal, and then tells them “now assuming 1+1=3, what does 2+2 equal” and then wondering over how they get different results from different people.


Chris 09.27.10 at 3:10 pm

I’m tempted to lecture on this just so I can use the pics.


novakant 09.27.10 at 4:10 pm

Ok, off the cuff I’ll try to apply this to a real world example:

3.8-5 million people died in the second Congo between 1998-2003 and nobody cared – so the international community basically walked away from the train tracks singing que sera sera. The switch would be equivalent to exerting international pressure to minimize the number of victims that would undoubtedly be and indeed have been killed. The fat man would be equivalent to military intervention, which would require us to actively kill people in order to save others.

The problem I have with utilitarianism is that, with hindsight (another problem of course), it could be used to justify us killing, say, 3 million people in a military intervention, because that would have been a better outcome, than the 3.8-5 million that actually got killed.


john c. halasz 09.27.10 at 4:13 pm


Well, there are actually four “payoffs” between the switch case and the fat man case: yes/no, yes/yes, no/no, and no/yes. IIRC there is supposed to be an empirical, experimental dimension, with respondents statistically choosing yes/no about 2/3 the time, yes/yes 1/6 the time, and no/no 1/9 the time, with no/yes statistically insignificant, except if your survey sample population is from Wall St.


qb 09.27.10 at 4:34 pm

novakant @ 107: The problem I have with utilitarianism is that, with hindsight (another problem of course), it could be used to justify us killing, say, 3 million people in a military intervention, because that would have been a better outcome, than the 3.8-5 million that actually got killed.

At least 800k-2 million people would want to know why you have a problem with that.


Earnest O'Nest 09.27.10 at 4:49 pm

109: As he started out by saying: he doesn’t care.


qb 09.27.10 at 4:58 pm

Um… no he didn’t. Or if he did I can’t find it anywhere.


Henri Vieuxtemps 09.27.10 at 5:07 pm

You can kill 3 million people in a military intervention, but how do you know that 5 million would’ve been killed without it? In the end, you’ve just killed 3 million people; how utilitarian is that?


bianca steele 09.27.10 at 5:20 pm

I wonder whether you would get different results if you primed subjects with the Milgram experiment, or with a discussion of how a “switch” connected to a computer-driven system works, or with a discussion of how a railway switch works, or with the fact that Hitler used trains.


Earnest O'Nest 09.27.10 at 5:35 pm

111: ‘nobody cared’.


Earnest O'Nest 09.27.10 at 5:40 pm

The international community is by the way doing an intervention in Congo. One that is heavily criticized for not intervening enough.

That being said, the international community did not intervene in Rwanda – and there are little people that would doubt that people were still very much killed.


geo 09.27.10 at 6:02 pm

ajay@105: Intriguing example. Of course Williams is right, just as Thersites was right, while Hal and Ulysses were lying upper-class bullies. Hard to know for sure, but I suspect Shakespeare agreed with Hal and Homer with Ulysses.


The Modesto Kid 09.27.10 at 6:43 pm

The Trolley Problem can only be enhanced by the presence of a spritely Sam-I-Am figure supplicating you to pull the lever. “You will not pull it. So you say. Try it! Try it! And you may. Try it and you may, I say!”


The Modesto Kid 09.27.10 at 6:46 pm

“Say! I like to pull that lever! Oh Sam-I-Am, you are so clever! I would pull it night and day. I would pull it June and May. I will pull it when I can. I will kill that tied-down man!”


roac 09.27.10 at 6:50 pm

Novokant @107: I am lamentably ignorant about conditions in the Congo in the period cited — do you have in mind some specific form of pressure that the “international community” could have exerted that would not have involved at least the threat of violent force? If there was such a lever, how would it have operated, other than by inducing the object of the pressure to send men with lethal weapons to the scene of the violence with authority to use them?

geo — Shakespeare’s personal views on any topic are notoriously inaccessible, since he was he had to make a wide variety of viewpoints dramatically plausible. But he could not have taken Williams’ side openly without the risk of putting his employers out of business. Thersites is in Shakespeare to BTW, in Troilus and Cressida:

I am a bastard too; I love bastards: I am a bastard
begot, bastard instructed, bastard in mind, bastard
in valour, in every thing illegitimate. One bear will
not bite another, and wherefore should one bastard?

(What Homer thought is doubly inaccessible as there is no agreement as to who he or she was or how many of him or them there were. The Singer of Tales, by Albert B. Lord, puts forward a fascinating theory that I find convincing as well.)


geo 09.27.10 at 7:23 pm

I am a bastard too; I love bastards: I am a bastard
begot, bastard instructed, bastard in mind, bastard
in valour, in every thing illegitimate.

Marvellous! My motto henceforth!


Tim Wilkinson 09.27.10 at 7:32 pm

ajay, geo: it’s an interesting appeal to double effect, because it’s undoubtedly the case that the slaughter of his own soldiers was in no way part of the king’s preferred plan – and indeed it was presumably not even foreseen – not with certainty, anyway.

Responsibility should thus be (even) less plausibly ascribed than in standard double-effect cases such as the DE defence of the fat man case. In those, whether the supposed ‘side-effect’ (death of fat man) can plausibly be claimed not to be part of the plan depends in part on how the concrete events should properly be described: stopping the train, putting a fat man in front of the train, dropping our corpulent friend 30 feet onto a track so that a 20-ton locomotive will be brought to a halt by his squished flesh, sacrificing his life to save others, etc. And ex hypothesi his death is expected with certainty.

But in the Shakespearean case, though the case for the king’s exculpation should if anything be stronger than it would be in standard double effect situations, I don’t think many people – even those sympathetic to DE – would really suggest that the king can use that kind of argument to avoid responsibility for the deaths of men he orders into battle.

Of course there is the novus actus interveniens side of things; there is more than one agent involved. The men could (a point Williams seems to finesse) mutiny or ‘dishonourably’ refuse to join up.

But responsibility is not zero-sum, so you may blame the men to some extent while still blaming their leader too, just as you might blame someone for ‘procuring’ an offence committed by someone else, or you might say it is someone’s own fault that they got attacked (they should have known better than to walk down there shouting that slogan, etc) without suggesting for a moment that their attacker is thus absolved from blame for attacking them.

(Earnest O’Nest: “there are little people that would doubt…”? Would one of those be Sam-I-Am by any chance?)


Earnest O'Nest 09.27.10 at 8:14 pm

122: I’m hugely ashamed that I have proven capable of being worse than ever.


Tim Wilkinson 09.27.10 at 8:38 pm

Yeah I thought it was (very, very) mildly amusing in an utterly lighthearted way at the time but after posting it I felt ashamed too. Unfortunately shame is not zero sum either, so we must both suffer these pangs.


qb 09.27.10 at 8:47 pm

113: Utilitarians are generally against killing 3 million people–unless the alternatives are even worse. How do we know the alternatives would be worse? In this case, novakant was stipulating the numbers. Obviously it’s harder to predict expected utility in the real world. But there’s a world of difference between acknowledging epistemic limitations on the one hand, and throwing up our hands and claiming, for instance, that military intervention is intrinsically morally wrong, on the other. Utilitarians can acknowledge that the uncertainty in calculating the moral costs of military intervention is a strong prima facie reason not to do it–but then, the certainty of civil war and genocide incurring very high moral costs is a strong prima facie moral reason to stop them, too.


geo 09.27.10 at 9:10 pm

roac: he had to make a wide variety of viewpoints dramatically plausible

Yes, but did he ever make the democratic/republican/dissenting/anticlerical viewpoint dramatically plausible? And of course he wouldn’t dream of displeasing his royal/noble employers and patrons — he never seems to have strayed, or been tempted to stray, a millimeter from the Elizabethan conventional wisdom.

Tim: I’m afraid I’m not at all inclined to exculpate kings. If they didn’t want to see their soldiers slaughtered, it’s only because they wanted to make use of them for their own frivolous and bloody purposes on another occasion. For the men’s actual welfare, kings typically didn’t care a pin. This is what (Homer’s) Thersites and Shakespeare’s Williams dimly grasped, and why we should sympathize with them against vain Hal and brutal Ulysses.


Dingbat 09.27.10 at 9:13 pm

Gahh! I shouldn’t bite, but I do, every time. What the Trolley Problem illustrates is just how far people will go to come up with a scenario where it’s okay to torture, kill horribly, maim, “put life at risk,” and have a “moral justification” for it. The Trolley Problem is itself the problem!

Yes, there are awful situations in the world where you have to choose from bad alternatives, and yes, you need to think about how to deal with those things, but to say that there’s some overarching philosophical framework that makes it ok, that will make it okay for the person who has to make this decision—especially when people who take philosophy classes is college (at least in the US) are the least likely to have to deal with situations like this–after having lived through a situation of awful alternatives, is, quite frankly, inhuman and absurd.

The Trolley Problem is a violent fantasy, and no more. It closes off thinking instead of opening it up, and what is more, it gives us an excuse to play with the life of a hypothetical fat person. Shoot, fat people have already committed a crime against society, haven’t they? They use too many resources and they’re ugly and they smell like fry oil. Shoot, they’re probably white trash, or black welfare queens, or, anyhow, they’re not like us.

If you think that ideas have consequences, how can you present this vignette to 18-year-olds and NOT expect them to be nudged a little closer to Not Caring About Their Fellow Man?


Tim Wilkinson 09.27.10 at 9:16 pm

No, I’m not that way inclined either.


Aulus Gellius 09.28.10 at 12:39 am

geo: last time we interacted in a CT comments section, as I recall, I was defending the greatness of Shakespeare against your criticisms. So I’m afraid I can’t resist jumping into the opposite position now. First of all, I think it’s often quite surprising to us moderns how much someone can seem to understand the unfairness and nastiness of lower-class social positions without ever reaching the step of, “and so that’s wrong.” Thersites’ speech in the Iliad is quite convincing, and the more so to readers/listeners who know that Ulysses is currently helping Agamemnon follow a deceptive dream precisely intended by Zeus to get a whole lot of Greeks killed and accomplish nothing. But I think it’s nevertheless clear that you’re supposed to think Thersites is a jerk who doesn’t know his place, and that it’s hilarious when Ulysses beats him up (the other soldiers certainly think so). It’s not that Homer doesn’t know how awful it is to have a bad commander, or be involved in a stupid war; but the only appropriate response is to shut up and hope things get better.

Similarly, Shakespeare can recognize that common soldiers are in an awful position beyond their control, but still think that that’s just the way it goes, not the king’s fault. In general, he’s a pretty solid believer that if the common folk start deciding things for themselves, they’ll just make a mess of everything: see Coriolanus, and whichever Part of Henry VI has Jack Cade in it. And if I remember right, in Henry V Williams ends up apologizing to Hal, and being graciously forgiven. Nor do I see much evidence for portraying this as a time-serving smokescreen to disguise Shakespeare’s deep faith in democracy and universal rights.


geo 09.28.10 at 1:25 am

Well-spoken, Aulus; herein hast thou seized
The very pith and marrow of the question!


novakant 09.28.10 at 1:30 am

I wasn’t saying that I had the only or right answer on what to do regarding the second Congo war, it’s a dilemma really, but the case poses interesting questions:

It was the bloodiest conflict since WW2, yet there was very little media coverage and it hardly penetrated the consciousness of the public at all. Why is that and, more generally, how do we choose the cases to which we apply our utilitarian calculations? How is it possible to be selective when we’re talking about the “greatest happiness to the greatest number of people”?

On utilitarianism in general: who appointed us to play god and bring about such a state, at what cost and what is the threshold where we become as evil as the enemy? Even if there was epistemic certainty that we would end up with more happiness and less misery if we took certain actions – what about “jus in bello”, how can you order people to “destroy the village in order to save it” and how can you look people in the eye when you’re doing it and tell them it’s for a greater good, you’ve done the calculations, they should understand and acquiesce?


John Holbo 09.28.10 at 1:30 am

Entertaining thread! I particularly like Modesto’s Seussisms.


Jim Rose 09.28.10 at 2:49 am

Do not aiplane pilots make this triage decision all the time when they try and crash planes in distress into empty fields or land them on highways rather than crash them into houses?

Flight deck tape recorders in black boxes show airline pilots matter of factly asking the co-pilot to find a ‘black spot’ in case they cannot make the nearest airfield during a night time emergency and they cannot otherwise identify where the houses below are.

what else would anyone expect them to do? Only philosophers can turn courage and grace under mortal peril into a moral dilemma for faculty debating clubs.


Substance McGravitas 09.28.10 at 2:55 am

Only philosophers can turn courage and grace under mortal peril into a moral dilemma for faculty debating clubs.



Oliver 09.28.10 at 6:39 am

The respondents throw the switch only because they have no further information why the people are tied to the tracks. If you tell them that the single man is the victim of an extortionist who holds him for ransom and the five people agreed to be tied as a party game, I bet they won’t throw the switch.
Likewise they’ll push the fat man, if you tell them that he tied the people to the track.
You must answer the question who rather deserves to die. Justice is a concept that cannot be explained by utilitarian means.


David 09.28.10 at 7:52 am

Suppose that instead of a fat man you had the Mona Lisa? Or, for a more suitable weight, the Venus de Milo, attached to a handy ramp The work of art would be completely destroyed.


john c. halasz 09.28.10 at 7:53 am

“Entertaining thread! “:

Ya. And for further entertainment “value”:


ajay 09.28.10 at 8:43 am

geo and others: I’d read that passage slightly differently. Williams isn’t saying “if we die in this battle, it’s the king’s fault, because we didn’t have a choice”. He’s saying “if we die badly in this battle, it’s the king’s fault”. Badly, in this case, means with unfinished business – either with responsibilities undischarged (widows and children) or with sins unforgiven. And, Williams adds, there are few die well that die in battle, because a battle doesn’t exactly give you much chance to sort matters out.

And Hal replies, essentially, “well, if you haven’t sorted your lives out before coming on campaign, it’s your problem, not mine”. That’s why the two examples are the son drowning “sinfully” and the merchant’s servant being killed by robbers “in many irreconcil’d iniquities” – and both, implicitly, being damned as a result of dying unshriven.

Sticking with the same play, incidentally, Shakespeare makes it very clear that Hal doesn’t have a choice in fighting Agincourt. He’s retreating from Harfleur to the British town of Calais and the French are in the way, deliberately forcing him to battle. The whole “anyone who doesn’t want to be here can leave now” bit in the “band of brothers” speech is irony – just as it would be if delivered by a more recent commander at Arnhem or the Chosin Reservoir.


ajay 09.28.10 at 8:48 am

Do not aiplane pilots make this triage decision all the time when they try and crash planes in distress into empty fields or land them on highways rather than crash them into houses?

No – think about the chance of the pilot and passengers surviving a belly landing on a large flat space rather than a landing that includes the aircraft smashing into a house at 90 knots.


John Holbo 09.28.10 at 9:48 am

“Ya. And for further entertainment “value”” … a book about genocide?

Sorry, john, is it your view that humor is, somehow, immoral? Or is it your view that genocide is ultimately just a joke?


ajay 09.28.10 at 9:58 am

Yes, but did he ever make the democratic/republican/dissenting/anticlerical viewpoint dramatically plausible?

In Act I scene 1 of Henry V itself, where Canterbury and Ely are plotting to distract the king from his plans to nationalise the monasteries (thanks Flanders & Swann) for the public good, by giving him an excuse for an enjoyable and profitable foreign war instead? Pretty cynical stuff, and kicks the props out from all the rest of the heroics in the play.


bianca steele 09.28.10 at 1:59 pm

TW @ 122:
By focusing on how things should be described, though, aren’t you shifting the focus to linguistic and social, historically conditioned concerns–thus doing something other than philosophy? The point of the exercise seems to be to identify relevant aspects of the problem for doing philosophy. This seems to explain why nobody in the Harvard video that I recall, or in any scholarly discussion of the problem I’ve seen mentioned, says something like that it is wrong to commit murder, or even that it is wrong to be violent.

I suppose it’s possible that people who might otherwise say “murder is wrong” usually feel one of the other alternatives is close enough to what they mean.


Njorl 09.28.10 at 2:02 pm

Have you considered applying for a grant to get real trolleys and volunteers so you can do a little experimental philosophy?


geo 09.28.10 at 4:08 pm

ajay @138: Yes, I think your reading is right. It’s not the soldiers’ mere lives but their estates and their salvation that the king is arguably endangering if “the cause be not good.” This doesn’t make me think any better of Hal (or Shakespeare), though. On the contrary: the right of the king to send soldiers to their death in a frivolous cause is, on this interpretation, now not even contested, as long as they’ve confessed their sins and left their families provided for. So now it seems that Shakespeare couldn’t even conceive of a popular challenge to divine right, even to swat it casually away.


ajay 09.28.10 at 4:21 pm

Fair point, geo. I don’t know enough about Shakespeare to know how the common soldier gets treated (or even mentioned) elsewhere…


loren 09.28.10 at 5:41 pm


qb 09.28.10 at 8:08 pm

novakant @ 131: I suppose no one “appointed us to play god,” but if your question is how anyone could have the arrogance to act in ways intended to result in the least amount of suffering–I guess I’d have to respond (in equally laden terms) by asking how much of a coward you’d have to be to see a path you have good reason to believe would prevent millions of deaths and then not take it for fear of appearing presumptuous. And were I to find myself in such an unfortunate situation, that’s exactly how I’d justify my actions to those whose lives would be forfeit. I would ask them: Would you have me allow the many to die, so that you, the few, may live? How could they look me in the eye and say Yes?


john c. halasz 09.28.10 at 8:45 pm


Midnight melancholia. It was a comment about the gap between a toy model and insoluble reality. (And it’s not a book about genocide; it’s about the Congo Wars, which novakant had mentioned above, an unfathomable, horrifying mess, in which all parties are derelict). Since the topic is moral judgment, which is about the (attribution and acceptance) of responsibility (and blame), and our evasive incapacity to bear them, eh?

Yes, it’s a tonally mixed thread, not unusually so, dependent on the contingent motives and pre-occupations of contributors. So, while we’re at it, I find geo’s pre-occupation with the (ir)responsibility of kings, which is not just anachronistic, but fictive, as well, rather silly.


Substance McGravitas 09.28.10 at 8:55 pm


Henri Vieuxtemps 09.28.10 at 9:23 pm

@147, but don’t you think there is clear, striking difference between the trolley-switch scenario and the scenario with the second track being free, empty – nobody dies when you flip the switch. Coward, really?


Henri Vieuxtemps 09.28.10 at 10:02 pm

…I mean, obviously arithmetic doesn’t work very well here. You saved one person and killed no one – that’s one life saved. You saved 5 and killed one – that’s 4 lives saved. 4 > 1. Yet the second equation is quite controversial, while the first one is completely unobjectionable. Why do you think that is?


qb 09.28.10 at 10:52 pm

Um… because not everyone is a utilitarian? I’m not sure I understand what you’re asking. Am I to refrain from doing what I think is right because other people disagree with me? As for the word “coward,” I was responding in kind to the equally overblown language of “playing God,” but since you’re taking issue with it, do you think fear of appearing presumptuous is not a cowardly reason to allow millions to die unnecessarily?


novakant 09.28.10 at 11:15 pm

qb, unless you’re a soldier who has actually been on the ground carrying out such orders, you have no right whatsoever calling people cowards – pushing a button, signing some order or pontificating on blogs doesn’t require any bravery, it’s very easy and that is part of the problem actually. And even if you were a soldier, I would question if in general it is braver to just obey orders, which is what soldiers are supposed to do and again part of the problem, or to disobey them, if your conscience compels you to.


Jim Rose 09.29.10 at 1:50 am

A concrete hypothetical:
• You are Truman’s chief defence counsel at his 1946 impeachment hearings charged with not using the atomic bomb as soon as possible to force the Japanese power elite to accept terms of surrender.

• Truman could have chosen to abjure from using the 2 bombs at his disposal and let the fire bombings burn down most Japanese cities and towns using the new air bases for B26s from Okinawa, let 100,000 Chinese be slaughtered on average every month at the hands of the occupying Japanese army, and invade in December and call forth a slaughter of a million or two more on both sides.

• The second bombing discredited the important faction within the Japanese ruling oligarchy that argued that the USA had only one bomb to use. (I think a third bomb may not have been available all that soon so good thing that this Japanese bluff was not repeated or many tens or hundreds of thousands more would have died in occupied China).

Would talking about the pros and cons the trolley problem helped Truman avoid a unanimous vote for conviction in the Senate?


Substance McGravitas 09.29.10 at 1:56 am

Those weren’t the only options for Truman.


qb 09.29.10 at 2:58 am

Oh please.


qb 09.29.10 at 3:04 am

Spare me the sanctimony, ffs.


Jim Rose 09.29.10 at 3:05 am

Substance McGravitas,

What were Truman’s other options? A demonstration explosion would have signaled to the Japanese oligarchy that the new president Truman is weak and reluctant to spill blood.

The willingness of the Japanese oligarchy to waste the blood of their own people and spill the blood of others without limit was central to their strategy of avoiding occupation and the dismantling of the old order.

Japan was governed largely by a consensus among an oligarchy of ruling factions. No major decisions of national policy could be reached until such a consensus had been obtained. This process inevitably took time and involved complicated pressures and struggles of will among those of differing opinions. Assassinations and the threat of the same often greased the wheels.

The Japanese oligarchy felt in the spring of 1944 that Japan was facing certain defeat or at least that the time had come for positive steps to end the war.

While Japan no longer had a realistic prospect of winning the war, Japan’s leaders believed they could make the cost of conquering Japan too high for the Allies to accept, leading to some sort of armistice rather than total defeat.

The Japanese army fought to the death with 99% plus casualty rates as the Americans moved from island to island to show that any attempt to invade Japan would be too high a price to pay. This was the Japanese ruling elite’s ace in the hole. The War Department staff in Washington estimated there would be 250,000 to 500,000 American casualties in an invasion of Japan.

Nearly 500,000 Purple Heart medals were manufactured in anticipation of the casualties resulting from the invasion of Japan. To the present date, all the American military casualties of the sixty years have not exceeded that number. In 2003, there were still 120,000 of these Purple Heart medals in stock!

After the bombings, a public admission of defeat by the responsible Japanese leaders was secured prior to an invasion and while Japan was still possessed of some 2,000,000 troops and over 9,000 planes in the home islands. There was no need for the Allies to either invade Japan or deal with the million Japanese troops in China who could have held out as a government-in-exile.


Substance McGravitas 09.29.10 at 3:27 am


geo 09.29.10 at 3:37 am

A demonstration explosion would have signaled to the Japanese oligarchy that the new president Truman is weak and reluctant to spill blood.

This is entirely speculative and sounds very dubious to me. The US had already inflicted horrendous damage on Japanese civilians through urban fire-bombing. The American blockade was constantly tightening; Japan was starved of essential supplies and its eventual defeat was obvious. The Emperor was solidly for ending the war, and the oligarchy was divided. An American demonstration, plus dropping the pointless demand for unconditional surrender, might very well have led to the defeat of the die-hards. This is certainly what Dower, Alperovitz, and other leading historians think. And it would have set a tremendous example of humaneness and rationality — it was a unique opportunity lost.

And of course, even more squalidly, there is certainly evidence that the US wanted to use the Bomb to influence the postwar settlement with the USSR.


Substance McGravitas 09.29.10 at 3:45 am

The weakness argument is fairly stupid given that the US was raining bombs on Japan at will throughout Truman’s tenure, continuing until the day of surrender.


Jim Rose 09.29.10 at 4:15 am

Thanks for the interesting link.

Your arguments are contradictory. Saying that fire bombings were just as ruthless as an atomic bomb, and then saying an atomic bomb was a higher stage of horror. Make up your mind.

You refer, for example, to Chester Nimitz, Commander in Chief of the Pacific Fleet, two months after the bombings:

“The Japanese had, in fact, already sued for peace before the atomic age was announced to the world with the destruction of Hiroshima and before the Russian entry into the war. . . .The atomic bomb played no decisive part, from a purely military standpoint, in the defeat of Japan”

What were the Japanese feelers?

Accept the Potsdam Declaration with the following conditions:

(1) That the Allied forces would not occupy the homeland;
(2) That the Japanese military and naval forces abroad would be withdrawn, disarmed and demobilized by Japan itself; and
(3) That all war crimes should be prosecuted by the Japanese Government.

Remember that one-half the imperial war council wanted to fight on after the two bombings.

There were coup attempts to stop the surrender with generals commanding Tokyo based troops holding back to see who ended-up on the winning side.

The willingness to sue for peace among the oligarchy was tactical and based on an armistice type deal to end to the war to live to fight another day.


Substance McGravitas 09.29.10 at 5:29 am

Your arguments are contradictory. Saying that fire bombings were just as ruthless as an atomic bomb, and then saying an atomic bomb was a higher stage of horror. Make up your mind.

I assume this is directed at geo, which is strange because it’s nothing like what he wrote. As to what Nimitz meant, I don’t know: I haven’t read his mind and won’t guess, but keep clicking away at the link provided and you’ll find some potential answers. I think you’ll find that your trolley requires a lot more maintenance than the one at the top of the thread.

There’s more at the top level of that site.


Jim Rose 09.29.10 at 6:44 am

Substance McGravitas,

name three Japanese peace feelers. what terms were sought?

The use of atomic bombs brought an end to a bloody war that would have become far bloodier if the planned invasion of the Japanese home islands had proven necessary.

The civilian peace party within the Japanese oligarchy had great difficulty prevailing over the hardliners even after both bombs had dropped and the Soviet Union had entered the war. remember the famous 3:3 vote. It was hardly a peace feeler.

The actual interrogation records of Japanese officials reveal that every Japanese official questioned but one (and he was contradictory) expected the war would have continued absent the shocks of the atomic bombs and Soviet entry.

Japanese leaders believed that if they could defeat or inflict terrible casualties on the initial American invasion, they could secure a negotiated end to the war to their satisfaction. Thanks to code-breaking, American leaders knew this.

The Japanese military was working around the clock to turn the site of the first planned invasion into massive killing grounds. Intelligence intercepts show this, and Japanese geography makes it very obvious were any invasion would have start.

No American president could have accepted the terms the Japanese oligarchy wanted such as no occupation and self-disarmament because this would have meant abandoning the Allies’ most basic war aims.

BTW, given the moral insights offered by the Trolley problem, what minimal terms of surrender would you regarded as acceptable from Hitler. The Japanese were keeping a good pace with Hitler on the slaughter of civilians in China and elsewhere in S.E. Asia . About 100,000 a month since 1937 in China; the massacre of Manila was 100,000 alone in one month in 1945 done by 10,000 Japanese marines in a rear-guard action.


Henri Vieuxtemps 09.29.10 at 7:30 am

@153, Um… because not everyone is a utilitarian?

Somehow I’m having difficulty trying to imagine a real-life person who would feel that killing one million people to save one million and one is exactly the same deal as saving one without killing anybody.

And if I’m right and such an individual doesn’t exist, then, I think, you’d have to admit that, even for the utilitarians, normal math doesn’t work here.


qb 09.29.10 at 8:51 am

Henri, I’m sorry, I’m even more in the dark now about your question than I was before. Why do you think utilitarians are committed to the claim that “killing one million people to save one million and one is exactly the same deal as saving one without killing anybody”?

I suspect you’re taking aim at versions of utilitarianism which claim actions are right just when they maximize average utility, but even if we restrict the discussion to those versions of the theory, utilitarians still have ample resources to explain why we ought to treat such cases differently in real-world situations, if not in the frictionless world of theory. I have difficulty trying to imagine a real-life person who believes that feathers fall as fast as coins, but if such people don’t really exist we need not conclude that an object’s gravitational acceleration must depend on its density.

But while physicists agree that gravitational acceleration is independent of density, not all moral philosophers agree–not even all utilitarians agree–that two scenarios with equal average utility are morally equivalent. The most dramatic differences appear when people claim that some kinds of action are intrinsically (that is, even in the frictionless world of theory) worse than others, e.g., killing versus letting die.

In your original question, you asked me why I thought that the rightness of saving one without killing anyone is uncontroversial, while the rightness of killing one to save five is controversial, and my perhaps overly glib response was that some people aren’t utilitarians. A better response, I think, would just be to say that the latter case is controversial because some people believe that killing is worse than letting die, and it is worse enough that it might be better to let more people die than to engage in killing, even in the frictionless world of theory.

I concede that there is disagreement about this point, but I fail to see why that disagreement is relevant to my response to novakant, both because I don’t see why the mere existence of disagreement about some moral dilemma counts for or against any particular solution, and because I don’t see how anyone could think that the act of killing is so wrong that it should prevent us from engaging in a military intervention which would certainly kill many, many people, but, it was stipulated, would ultimately result in 800k-2 million fewer casualties.


Jim Rose 09.29.10 at 10:11 am


How is the trolley problem different from the valuation of a statistical life issue in cost-benefit analysis?

For more than a quarter century, economists have developed empirical estimates of the trade-off between wages and fatality risks to assess the benefits of from health, safety, and environmental regulations to road building. These benefits of saving a statistical life values are critical inputs to the policy because the benefits from reducing risks to life are often the dominant benefit component.

In the late 1970s, Gordon Tullock wrote a book review about avoiding difficult choices. We make a decision about how to allocate resources, how to distribute the resources, and then how to think about the previous two choices.

People do not want to face up to the fact resources are scarce and they face limits on their powers.

To reduce the personal distress, Tullock observed that people often allocate and distribute resources in a different way so as to better conceal from themselves the unhappy choices they had to make even if this means the recipients of these choices are worse off and more lives are lost than if more open and honest choices were made up about there only being so much that can be done.

The trolley problem spends a lot of time pontificating about how we must make choices with tragic consequences rather than give guidance on what must be done because scarcity of resources requires the valuation of life in everything from health, safety, and environmental regulations to road building. health budgeting is full of tragic choices about how much is spend to save so lives and where and for how long.


Guido Nius 09.29.10 at 11:23 am

I think Jim is entirely right. The trolley problem is irrelevant not because it is so uncommon but because some trade-off is always inherent in what we do. The choice not to go to war is as much a choice as the choice to go to war. The non-consequentialist part of utilitarianism is what we’re choosing as a good (to be maximized) outcome. Take treating Alzheimer – if maximizing quality of life is taken as to be maximized we’ll come to other conclusions (including euthanasia) than if we want to maximize life full stop. There is no utilitarian’solution without some non-utilititarian assumptions.


Henri Vieuxtemps 09.29.10 at 11:49 am

I don’t see how anyone could think that the act of killing is so wrong that it should prevent us from engaging in a military intervention which would certainly kill many, many people, but, it was stipulated, would ultimately result in 800k-2 million fewer casualties.

And I’m saying that “800k-2 million fewer casualties” is not necessarily a convincing justification (even for an utilitarian; not an ideal, but a real-life one), if the trade-off is killing 3 million people.

Suppose you can prevent a bloodshed in Congo, that you know will kill 5 million people – by killing 3 million people in, say, China. Or in the city where you live, for that matter. Is it obvious to you that it would be a right thing to do?


Jim Rose 09.29.10 at 12:12 pm

Guido Nius, thanks

Another daily example is ambulance drivers.

In a court near me, an ambulance driver was charged with causing death by negligent driving. The offence carried a maximum of three months in the big house.

A little old lady stepped out in front of his ambulance while he was rushing to an emergency.

The ambulance driver was driving at the maximum speed of 90 km that he was permitted to travel while on emergencies on urban streets with the sirens are on. The little old lady who died after stepped out in front of his ambulance was deaf.

The jury took one look at the ambulance driver in the dock in his ambulance uniform and said to themselves their but for the grace of god goes the man who will rescue me or my family in an emergency and found him not guilty.

The ambulance driver had the full support of his employer, was not under suspension, and wore his uniform in the dock with the permission of his employer, but was desk bound for the duration of the legal proceedings.

The trolley problem did not tie up any of the random members of the public that make up juries in their jury deliberations or cause even one of them to say guilty.

The trolley problem arises every time an ambulance or fire truck speeds to an emergency. The truck speeds up but within a balancing of the very real risks to life of both delay and speed.


Tim Wilkinson 09.29.10 at 1:28 pm

bianca steele @142 – In operational terms, for double effect justification to work you have to come up with some canonical description of your plan, and make it plausible that you could wish essentially that plan to be executed without the undesirable (‘side’) effect happening.

How it should be described = how one would describe it properly, correctly for these purposes. If you describe the relevant action as diverting the train away from its potential victims, then double effect might conceivably be invoked to justify the switch-flicking. If you describe it as diverting the train onto the other track where one person is tied, probably not. So asking which description is correct for these purposes is one way of putting the issue.

If you don’t think there is a correct description (determinate nature, whatever) of the plan, then one could always wish that by some miracle or other transfiguration of reality, the bad effect wouldn’t come about (The fat man would stop the train but not die). If you have to come up with some canonical (correct) description, there may be a non-trivial answer to the question whether double effect justification is available. I don’t know how this would be done (I don’t know what kind of counterfactuals you would use to establish that the bad consequence is inessential to the plan), but at least there is prima facie some hope of doing so.

(I don’t believe in double effect justifications for reasons related to this, besides other much more general reasons).

Droning on a bit here I fear, with little progress towards clarity. I’ll try to conclude: it’s true that the question is not essentially about descriptions. It’s just that you always end up using descriptions to, er, describe how things are to be conceptualised, what their essential character is for a given purpose, what sense we mean our terms to have etc. So philosophy types often end up talking about ‘under what description’ something is believed, planned, intended etc. I think that usage may have fallen off since its heyday in mid-last-century, though. But basically, talking about the correct description is a way of operationalising the problem for language-bound purposes like debate.


Substance McGravitas 09.29.10 at 2:02 pm

name three Japanese peace feelers. what terms were sought?

Oh, now you’re looking for three. That’s a tough trolley to keep in service! Well then, go look. Again, Truman had options, and if his own military staff figured the war had been won, I’ll go with them over Jim Rose.


Tim Wilkinson 09.29.10 at 3:20 pm

I’m far from being any kind of expert, but had the impression that the bombs were the first blow in the new Truman administration’s (resumption of the) Cold War. They also (no doubt as a side-effect) functioned as excellent tests of the effects of those bombs when dropped on cities.

Even if it’s true that the Japanese were not ready to surrender, there seems to be a bit of a gap between that and ‘A-bombing two large civilian population centres was the thing to do’.

I’m also especially impressed with the obfuscatory claim that ‘the bombings and Soviet entry’ were responsible for getting the Japanese surrender, as if they formed a single phenomenon.


burritoboy 09.29.10 at 4:28 pm

Returning to the initial point John was making:

This type of thought experiment within Anglo-American analytical philosophy could be described as a form of literary genre (just as dialogues, treatises, lives of the philosophers, philosophic novels, the philosophic letter, mirrors of princes, etc are other philosophic literary genres). The resulting question is what does this genre do within philosophy? How does this genre operate compared to other genres?

At it’s heart, this is a discussion of what philosophy is. The genre utilized is a reflection of how the writer (or readers, for that matter) understand how knowledge operates. Plato and Xenophon both write Socratic dialogues, but their dialogues (in my analysis) are quite different because Plato and Xenophon have a disagreement on what knowledge is and how best to gain it.

The resulting question from the above is: how does this genre (analytic thought experiments) encode what understanding of knowledge?

Here’s just some random thoughts about this:

That this genre is an entirely abstract thought-experiment encodes an understanding that non-philosophers (undergraduate students) can easily (with little introduction) and universally engage in very high-level abstract thought. This may be true, but we note that most of the other philosophic genres encode a contrary understanding – the Socratic dialogue, for example, asserts the contrary. Socrates (usually) encounters a named person in the agora, usually not one eager to philosophize, and only gradually moves to abstraction, often starting from a very concrete problem of the interlocutor’s (Alcibiades wants glory, Callias wants to gain the love of Autolycus, etc). Within the dialogue, Socrates will develop his arguments differently depending on the interlocutor’s intellectual ability, experience with philosophy, knowledge bases, character and so on. Socrates explicitly does not have a universal method to start philosophizing – the early writers of dialogues (Plato, Xenophon, Aeschines of Sphettus) all seem to agree that each individual interlocutor has his own affinity for philosophizing and these affinities can be radically different. Most of the interlocutors do not believe philosophizing will aid them in their aims and quite a few are extremely hostile to it.

This genre also has a specific literary style that attracts certain people and repulses others. This style especially rewards participants who like and excel at certain forms of argumentation. The genre asserts that philosophy is like this (analytic thought experiments) and those who do this easily are the most promising potential philosophers. Again, while this may be true, many other philosophic genres disagree. It’s notable that analytic Anglo-American philosophy uses extremely few literary genres compared to other philosophic schools – the early Socratics used dialogues, treatises, philosophic letters, autobiography (Xenophon’s Anabasis), novels (Xenophon’s Education of Cyrus) and other genres, for example. Rousseau deploys many genres – autobiography, treatise, dialogue and reverie among others.


geo 09.29.10 at 4:33 pm


Your arguments prove that tough negotiations, with a credible threat to use escalating force (including passive force, like a blockade), would have been necessary to end the war without using atomic weapons, and even then might not have secured all the peace conditions the Americans sought. So what? Was this too much to ask for the sake of not using a weapon that everyone, even most military commanders, foresaw would be a permanent blight on the human prospect?

The “most basic Allied war aims” were not sacrosanct. The proper aim of war is self-defense. Insistence on forcible regime change for sake of preventing possible future aggression is more problematic than you seem to recognize. And the non-explicit American war aims — to humiliate the Japanese by insisting on the right to depose the Emperor and to show the Soviets that we had a super-weapon and no scruples about using it — were not merely problematic but contemptible.

I realize that the above doesn’t settle the matter. But perhaps we’ve strayed too far from the post’s original philosophical topic?


novakant 09.29.10 at 4:47 pm

I think it’s essential to also look at the Transplant Problem – we don’t kill people to harvest their organs and I don’t know anybody arguing we should do so, even though it would make sense from a utilitarian point of view.


bianca steele 09.29.10 at 4:54 pm

Tim Wilkinson:
I’m not myself opposed to thinking in terms of things like double effect. I guess I am thinking that what John Holbo says in the last paragraph of @58–assuming he’s talking about the contents of the problem and not the seminar situation in which they’re typically discussed, which I’m not sure about–may actually be inherent in the “genre” of thought experiment, the way they are actually used. That is, the descriptions can’t be “thicker” without changing what’s being discussed. They have to be abstract and “thin” or they don’t map onto the details of the thought experiment in the way you need them to.

(Of course, the experiment is always described in words, so maybe they are never perfectly paper-thin and transparent. For example, it just occurred to me that “fat man” was the code name of the bomb dropped on Hiroshima–though Harvard undergraduates are probably mostly more sophisticated than to be misled by that.)


bianca steele 09.29.10 at 5:31 pm

Also, I would wager that a decent proportion of the students have seen similar discussions elsewhere, and associate the “strongest utilitarian” version with Communism, to which they are vehemently opposed. So they are probably going to favor the deontological theory (and, I’d guess, will assume it does include traditional morality like “don’t murder, no matter what” without modification).


Tim Wilkinson 09.29.10 at 5:31 pm

At it’s heart, this is a discussion of what philosophy is.

But how should such a discussion proceed?


burritoboy 09.29.10 at 5:48 pm

“But how should such a discussion proceed?”

I can’t say I know. My own take is above – the literary genre of a philosophical piece is itself a crucial part of the philosophic content of the piece. Analytical Anglo-Americans tend to use an extremely limited range of philosophic genres (compared to the history of philosophy, which has included a wild variety of genres) and the thought experiment is one within that extremely limited range. That might mean that comparative examination of genre might be illuminating – I attempt an initial stab above.

John’s mention of the contextuality of genre is also interesting. Something that could be very useful when students have reached the seminar room stage might however completely backfire when students are in earlier stages.

What I don’t think is that particularly useful is playing around inside the thought experiment itself – the internal content of the thought experiment was peripheral to John’s initial post.


loren 09.29.10 at 5:48 pm

Jim: “The trolley problem spends a lot of time pontificating”

Um, the problem isn’t pontificating (in either plausible reading of that utterance).


Guido Nius 09.29.10 at 6:46 pm

177: some people do kill people to harvest organs and there is no utilitarian defense of it, unless you want to make utilitarianism about personal profit – also, we do harvest organs and there are such good reasons for it that many people believe we ought to do it by default whenever there is a corpse that has potential in that regard.

I mean, you really insist on simplifying things to the point that you can blame the others and not run any risk to be blamed yourself.


Henri Vieuxtemps 09.29.10 at 6:58 pm

I thought Kevorkian was killing people to harvest organs, and indeed on pure utilitarian grounds.


Guido Nius 09.29.10 at 7:48 pm

You have access to information I haven’t. In any case, if so: on ‘allegedly’ utilitarian grounds is probably more accurate.

By the way: I am pro-euthanasia and I am also pro-default organ donorship so as long as there’s due process avoiding the organs being a reason for the euthanasia, I think it is common sense to make sure that the organs of somebody that has freely chosen to die are used for those that are freely choosing to live.


Henri Vieuxtemps 09.29.10 at 8:54 pm

I probably got it from the movie about him. But here’s a quote from the Frontline website:

That year, Kevorkian’s article, “The Last Fearsome Taboo: Medical Aspects of Planned death,” appeared in Medicine and Law. He argued that modern society, facing the longstanding prohibition against planned death, was “subjected to unrelenting paternalistic control based on moral codes that are rapidly becoming obsolete.” He called the medical profession “physically (philosophically) retarded, drifting aimlessly without a coherent or even workable ethical code.”

Kevorkian outlined his vision thus: “The acceptance of the planned death implies the establishment of well-staffed and well-organized medical clinics (‘obitoria’) where terminally ill patients can opt for death under controlled circumstances of compassion and decorum….Physician involvement should extend far beyond mere termination of life to permit exploitation of the enormous potential benefit that could accrue from the acquisition of organs for transplantation and the performance of daring and otherwise impossible human experiments under irreversible general anesthesia.”


qb 09.29.10 at 8:55 pm

Yes, the transplant problem is an excellent way to generate or uncover deontological intuitions; when I introduce the kinds of problems ethical theory attempts to address in my classes, I often juxtapose the transplant case with other cases intended to draw out utilitarian intuitions (I don’t generally use the trolley problem, though it would work just as well). How are we to square the apparently obvious need to make sacrifices for the greater good with the equally apparently obvious need to refrain from certain kinds of paradigmatically wrong actions, like killing? To the extent that you do not roll your eyes at the attempt to achieve theoretical consistency on these matters, or simply shrug and say “I’d rather do economics,” or some other field, you’ve begun to do ethical theory.

Utilitarian responses to the transplant problem usually resort to the notion of rules or heuristics to avoid the conclusion that we should be killing people to harvest their organs–in general, they will argue, a policy of allowing such transplants will have negative effects on overall or average utility for empirical reasons, even though there may be rare cases in which violating that policy will in fact maximize utility. To be sure, utilitarians need to bite the bullet: sometimes uncomfortable sacrifices will need to be made, and so much the worse for our deontological intuitions. The problem is that deontologists have their own bullets to bite–which, as in the trolley case, will usually involve allowing whole gaggles of people to die for the sake of keeping one’s hands clean–so it’s not as if the counterintuitive aspects of utilitarianism necessarily render it inferior to its competitors.


Jim Rose 09.29.10 at 9:18 pm

Substance McGravitas,

I see that you are unwilling to discuss here and directly what terms these peace feelers sought. I assunme it is because the terms sought were never a starter.

Intellegence intercepts and post-war interviews of the Japanese show that the japanese were committed the operation ketsugo until the bombing made this a futuile endeavour.

a better source of japanese intentions are the japanese themselves.


thanks for the post, yes, perhaps we are straying from the topic.

I put up Trumen’s 1946 hypothetical impeachment hearings for not going forward with the atomic bombings as a practical example of a trolley driver with no easy choices.

many of the responses arguedthat the trolley have additional tracks to divert too rather than argue what is to be done if the commitment of the Japanese to operation ketsugo was real and credible. This is an examaple of Tullock’s avoiding difficult decisions behaviour.

Would more have died from a starvation blockage and fire-bombings than from the use the bomb? why should so many more die to avoid a difficult decision that shorterned the war?

the trolley problem is about avoiding difficult decisions.


Substance McGravitas 09.29.10 at 9:22 pm

I see that you are unwilling to discuss here and directly what terms these peace feelers sought. I assunme it is because the terms sought were never a starter.

Fair enough, your arguments rest on assumptions so one more won’t hurt.


Jim Rose 09.29.10 at 9:46 pm

Substance McGravitas,

if the cabinet tried to surrender, what would have prevented a successful military coup? plenty of generals sat on the fence during the attempted coup of 12-15 August 1945.

are far easier device would have been for the navy or army minister to resign. this resignation would have brought down the government and ended any real peace feelers not desired by the military

what terms did the Japanese want in the peace feelers?


Substance McGravitas 09.29.10 at 10:19 pm

if the cabinet tried to surrender, what would have prevented a successful military coup? plenty of generals sat on the fence during the attempted coup of 12-15 August 1945.

An unsuccessful military coup is evidence for a successful one. Gotcha. In any case, What Jim Rose Assumes The Japanese Would Have Done has no bearing on whether or not Truman had other options, which he did, nor does it speak to what Truman thought the Japanese were up to.


Jim Rose 09.29.10 at 10:27 pm

Substance McGravitas,

another example of the trolley problm is peace feelers from Hilter.

the trade-off is less military deaths but the gas chambers stay in operation.

what would you have done if Hitler made peace feelers?


Tim Wilkinson 09.29.10 at 11:02 pm

many of the responses argued that the trolley have additional tracks to divert too rather than argue what is to be done if the commitment of the Japanese to operation ketsugo was real and credible.

Oh, so are you just using the actual events as inspiration for a stipulative fictitious example then? Some of your remarks didn’t make that entirely clear. But please do go ahead and lay out the precise formulation of the exhaustive set of alternatives with their complete consequences.


Tim Wilkinson 09.29.10 at 11:12 pm

Back in the real world (should you choose to go there), there’s a bit of a gap between a commitment of ‘the Japanese’ to Operation Ketsu-Go and the best thing to do being A-bombing a couple of cities.


Jim Rose 09.29.10 at 11:27 pm

Tim Wilkinson ,

The historian Chalmers Johnson has written that:

“It may be pointless to try to establish which World War Two Axis aggressor, Germany or Japan, was the more brutal. The Germans killed six million Jews and 20 million Russians [i.e. Soviet citizens]; the Japanese slaughtered as many as 30 million Filipinos, Malays, Vietnamese, Cambodians, Indonesians and Burmese, at least 23 million of them ethnic Chinese. Both nations looted the countries they conquered on a monumental scale, though Japan plundered more, over a longer period, than the Nazis. Both conquerors enslaved millions and exploited them as forced labourers …”

The trolley problem with either Hitler or the Japanese military is many more civilians would die if the war was not ended as qucikly as possible with atomic weapons.

Would the japanese ruling elite have surrendered in August 1945 but for the atomic bombings?


Substance McGravitas 09.30.10 at 1:09 am

Jim, the thing about the trolley problem is its cleanliness: the complications are brought in by the people it’s presented to. The thing about WWII is that it’s more complicated than you think it is.


Jim Rose 09.30.10 at 1:20 am

Substance McGravitas,

good point, but the person driving the trolley is human.

As Tullock argues, to reduce the personal distress of making difficult decisions, people often allocate and distribute resources in a different way so as to better conceal from themselves the unhappy choices they had to make even if this means more lives are lost than if more open and honest choices were made up about there only being so much that can be done.

a major driver of the opposition to the use of cost-benefit analysis and the valuation of statistical lives is its adoption makes people confront the tragic consequence of any of the choices available to them.

by saying how dare you value a statistical life does not change the fact that choices made without this knowledge will still have tragic consequnces and more lives may be lost because people want to conceal from themselves the difficult choices that they are making about others as voters and as policy-makers.

As soon as the trolley is real and the tracks take you to either deaths in Japan or even more deaths in the Japanese occupied countries, people grab at every possible straw for evidence that there is a third track.


novakant 09.30.10 at 1:28 am

Alright, qb, you are the brave philosopher prepared to kill, maim and march through rivers of blood in order to maximize utility and achieve consistency, while I’m the feeble minded ingenue, desperately clinging to outdated deontological intuitions such as “first, do no harm” and “though shalt not kill”.

btw, can we have your liver, then?


Jim Rose 09.30.10 at 1:42 am


would you have opposed the use of violence to depose Hitler? Were his would be assassins immoral?


Tim Wilkinson 09.30.10 at 2:14 am

The trolley problem with either Hitler or the Japanese military is many more civilians would die if the war was not ended as qucikly as possible with atomic weapons.

Well, if we are back in the real world again, one question would be: would more die under any alternative strategy (e.g. only Hiroshima A-bombed; military targets A-bombed – not that I suggest either of these would indeed have been best) than actually did?

There is perhaps the Cold War nuclear arms race to take into account, too. Maybe this was not a consequence, or maybe it would be a benefit, as making sure the Cold War was as cold as it was? Epistemic issues become particularly clearly relevant when we are talking about such remote consequences , as do the ethical aspects of willing or reckless ignorance, etc. Perhaps we also look at the actual or expected consequences of the use of such illegal WMD; perhaps Stalin’s reaction, US military hegemony, I don’t know. Should the bombs have been used earlier, perhaps? Should Roosevelt be held responsible for net deaths resulting from his forebearance?

These kind of attempts at calculation are too hard, even assuming we have complete info about the state of the world at t + the deterministic laws of nature. If we have actual or epistemic indeterminism (and we do have the latter), they are probably NP-hard, though with an input string like that, I should think polynomial time is still prohibitively long.

Would the japanese ruling elite have surrendered in August 1945 but for the atomic bombings?

I assume that’s intended as a rhetorical question – I certainly had the impression that you were (when in real -world mode) actually arguing positively not only that the actual bombings were the best of all available courses of action (maybe we relativise this to the information available to, or actually possessed by, the US command, and the consequences that were or should have been expected…), but that their being necessary for the surrender you speak of being no later than it was is a lemma in the argument explaining why that’s the case.

Personally, I’d have to look it up.

Independently of that issue: I’m far from being any kind of expert, but I had the impression that the bombs were the first blow in the new Truman administration’s (resumption of the) Cold War. Maybe not, or maybe that was just a handy side-effect of the best humanitarian option.


john c. halasz 09.30.10 at 2:27 am

Do facts count here? Livers can regenerate, and live liver donations are possible, originally only with pediatric cases, but nowadays in adult cases. Not that there are no risks and controversies, but it’s empirically not an either/or instance.

Incidentally, novakant, you did seem to say or imply something interesting in the Congo Wars case: IFAICT you offered a consequentialist reason for a de-ontic standard.


geo 09.30.10 at 2:55 am

Jim: parenthetical question — where’s that Chalmers Johnson quote from?


Jim Rose 09.30.10 at 2:57 am

Tim Wilkinson, thanks for the comments

I always thought references to the russian invasion as the decider was really desparate. Losing Tokyo in one atomic bombing is less important than losing Manchuria?!

after world war 1, the germans invented all sorts of fantasies about what really happened to excuse the fact that they sued for peace because they were routed.

How many more Chinese would have had to die as the Russians advanced through Manchuria in late 1945 so that the handwringers past and present could sleep better at night and avoid having to face up to some difficult choices?

Your link shows that the bombing shortened the war by several months at the minimum.

This results holds even using after paul nitze’s selective quotation of one Japanese high-level source in the strategic bombing survey that the war might have been over by Christmas.

December is not August.

The strategic bombing survey is a bit light the risks of a successful coup, and it mentions but does not explore in proper detail the capacity of the navy or army minister to resign and bring down a government at any time in august 1945 as did happen to the Tojo government after it has served its purpose to the ruling elite.

a word search of the the link you supplied does not find the word coup. The failed coup was after much reflection and temporising by the military high command and joining the plot or organising their own coup see

In any case, the actual interrogation records of Japanese officials reveal their statements that the war would be over by christmas or november were literally the reverse of Nitze’s assertions.

the atomic genie was out of the bottle by 1945. The only way to put the genie back in, as Tom Schelling observed in 1961, is universal brain surgery to eliminate all knowledge of both atomic weapons and the principles of how to make them again.


novakant 09.30.10 at 5:06 am

Jim, I’m sure you will be delighted to hear that I am not a pacifist and would have put a bullet through Hitler’s head, if it would have changed anything for the better. You will probably not be so happy to hear that I can think of several Anglo-American politicians, who would have suffered the same fate at my hands, if that would have contributed to the end of the mass slaughter they caused. That said, I’m not quite as nonchalant about killing large numbers of innocent or even not so innocent people as you seem to be.


Jim Rose 09.30.10 at 6:49 am


The trolley problem is about have a set of choices where all have tragic consequences. There is no third track that allows the driver to escape without blood on his hands.

History is full of examples of well-meaning attempts to regulate wars by prohibiting classes of weapons. The 1920s is an example of the dangers of such self-delusion about there is a third track or a slower trolley out there in international relations.

The Washington Naval Treaty of 1922 was designed to stop the naval arms race that helped to precipitate World War I. It limited naval tonnage among the five principal powers of the day. Germany was constrained by the Treaty of Versailles. The London Naval Treaties of 1930 and 1936 followed.

Japan, Italy and Germany ignored their treaty obligations, and the democracies limiting only themselves – reducing their ability to defend themselves. Germany was constrained by Versailles from building an air force – a fact of no comfort to the civilian victims of the Luftwaffe in World War II.

The 1928 Kellogg–Briand Pact (formal name: General Treaty for the Renunciation of War) was signed by the USA, France, Britain, Germany, Italy, Japan and a great many other states. The pact renounced aggressive war, prohibiting the use of war as an instrument of national policy except in matters of self-defence.

Obviously, the Kellogg–Briand Pact did not live up to its aim of ending war. This Pact did serve as the legal basis for charges of crimes against peace charges.

War criminals such as Bill Clinton and most heads of government in the EU in the late 1990s should be subject to those citizen’s arrests that so called peace activists try so as to make a spectacle of them-selves because the aforementioned war criminals supported the bombing of Serbia without a UN authorisation.

Go back to psychological crutchs, what does the trolley problem have to say about NATO’s unlawful intervention in the Kosovo war? Civilians died due to the NATO bombings.

Is the effect of a well-meaning but ill-considered effort to make humanitarian intervention obligatory as a matter of international law yesterday Kosovo, today Iraq, and tomorrow Darfur?


Guido Nius 09.30.10 at 7:11 am

186- Cool!


Henri Vieuxtemps 09.30.10 at 8:51 am

Yeah, Kevorkian is a true utilitarian. Apparently his other idee fixe was experimenting on, and harvesting organs from death-row convicts. It’s actually kind of funny, in a grim sort of way:

Professional journals and popular magazines rejected Kevorkian’s article outlining his proposal [to experiment on death-row prisoners]. Kevorkian later admitted his essay “reeked with sophomoric idealism and was highly impractical.” But he was able to present it at the December 1958 meeting of the American Association for the Advancement of Science in Washington. D.C.– only because the group’s chairman figured the outlandish scheme would turn more people against capital punishment. Kevorkian’s speech caused a minor stir in the press, but drew no official support save for a stirring endorsement from an animals rights group, which pointed out that cutting up humans would save the lives of rats and guinea pigs.


Jack Strocchi 09.30.10 at 9:07 am

John Hobo said:

Utilitarianism and its discontents. I also think the bug – namely, these situations are silly – can be a feature. You pump students for their ‘intutions’. Then you talk about how making the stories silly is likely to distort responses.

Philosophers, following the Sophists, have been trying to play “gotcha” to Kings since Plato was a lad. Its one way to improve your status, although it is notable that the Socrates, no slouch in the grey matter department, thought it best to follow the rules even when it meant putting his own head on the chopping block.

Finding fault with utilitarianism is an interesting intellectual exercise but there are major provisos. One is that “utilitarianism” is not a one-dimensional straw man for Kantians on a “search and destroy” mission. Another is that “intuition” does not always pump truly, indeed it evolves like everything else.

I think I can save a form of utilitarianism from an outbreak of “Trolleyological” hand-wringing by allowing Machiavellian act-utilitarianism as (literally) an exception to Millian rule-utilitarianism. Hopefully this goes some way to clear up some of the alleged conflict between Good acts and Right rules, at least for the common-or-garden utilitiarian.

Its not in dispute that there are two forms of utilitarianism:

Act-utilitarianism: For individual autonomies in full posession of the facts, can know the consequences and be prepared to take full responsibility for their deeds.

Rule-utilititarianism: For institutional authorities that do not have all the facts of each case fully in mind, and instead have standard procedures to deal with numerous like-cases.

We can all think of utilitarian acts which are intellectually defensible but instinctively repulsive eg trolley homicides, organ harvesting. Thats because in our thought-experiments the moral problem is simple and constrained, so we can “play God” and pick winners. Its just a matter of toting up all the scores.

But real life is not a thought-experiment conducted with Olympian detachment by philosopher kings. Its complex, uncertain and played out in real time. We have to make tragic choices, so to economise on unease and uncertainty we play by the rules.

Kant and Mill both agree we should follow society’s rules and should not second guess them. According to Hayek we are wise to all follow utilitarian rule evolution because “who died and selected me as Pope?” Over time these rules become second-nature to right-thinking people, which is how Good acts eventually get formalised into Right rules.

But there are sometimes hard cases (usually in war-time, surgery or other life-or-death “trolley” cases) where rule-utilitarianism can and should be set aside and replaced by act-utilitiarianism. Generally, hard cases make bad laws.

But exceptions can always be found. The proviso is the rule-breaker knows what they are doing and is prepared to accept moral responsibility for his violent acts. Thats when the individual autonomy can literally play God and usurp institutional authority. I dub this the “Machiavellian moment” where a Good act (end) can justify breaking the Right rule (means) . I daresay Truman did not savour it with relish.

In fact one CT blogger has already resolved the conflict between utilitarian acts and utilitarian rules in the classic case of “ticking time bomb” torture scenario. Pr Q’s argument implies that an individual official could carry out a (utilitarian) ethical act of torture provided this was done as an act of civil disobedience against generally-0bserved institutional rules against torture.

The torturer could then turn himself in to the authorities and reasonably hope for an acquittal if the act of torture did in fact save more lives. If not then the torturer would have to cop it sweet. Either way, the institutional rule against torture is still generally respected as Right. Whilst we may allow that the individual act of torture did some particular Good.

These fanciful thought-experiments make us queasy, because our instinctual notions of what is Right can often conflict with the intellectual pursuit of some Good. Of course unconscious evolution is not always Right, which is why a little conscious revolution is a Good thing, now and again.


Jack Strocchi 09.30.10 at 10:40 am

Substance McGravitas @ #173 said:

Truman had options, and if his own military staff figured the war had been won, I’ll go with them over Jim Rose.

Speculating on the the figurings of American strategists or the imminence of Nipponese surrender and $2.50 will get you a cup of coffee. No one doubts that “Truman had options”, but the best one was to end the war sooner rather than later. That is the option he chose for thos reasons and it was the right one, on utilitarian grounds alone.

We know what the Nipponese authorities thought. We have it from the horses mouth (the Emperor) that the dropping of the A-Bomb caused them war to end with unconditional surrender. No one has given a plausible explanation of why we should not take his words at face value. Its not as if he just made it up as he went along.

What the American military staff may have privately thought, or even publicly recollected in tranquility, is something for psychoanalysts and biographers to quibble over. What they actually did is a matter of historical fact: plan and prepare for a six-to-twelve month conventional military campaign, with expected military and civilian casualties well north of one million.

Thats several times more civilians that were killed in Hiroshima/Nagasaki. So from an act-utilitarian perspective the dropping of the atomic bombs is a no-brainer: bombs away! Although its not the kind of military act that I would want to see made into a rule of thumb.

The idea that the JIA was on the verge of surrender in August 1945, prior to acquitting itself in an Armageddon battle against either the Red Army or US Army, is risible. Bushido honour code demanded that oceans of blood be spilt. Okinawa was just an overture. The only question remaining was the eternal one: Whose by Whom?

On the strategic question many critics of the A-bomb have pointed to the successful Red Army attack on the Japanese Imperial Army in Manchuko and the imminent Soviet invasion of the outerlying Japanese islands as an argument against the A-bombing. Supposedly the Red Army would have cleaned up the Nipponese remnants without the need for “that awful thing”.

To me the Red Army advance was an obvious argument in favour of dropping the A-bomb. Having the Reds divide up Japan, as they divided up Germany and Korea, would have been a recipe for unending geo-political pain (and another possible nuclear flashpoint). As it is the successful use of the A-bomb stopped the Red Army in its tracks in North Korea. That surely made the South Koreans happy.

More generally, or perhaps particularly, I have to wonder about the rarefied discussions of morality by people who were not, and never will be, on the firing line. (Although lest we forget, the A-bomb was Einstein’s idea and he was a pacifist.) How many CT readers and writers are willing to martyr themselves (or loved ones) in an agonizing death on behalf of a contentious philosophical idea. Yeah, me neither.

Its significant that combat veterans of some philosophical sophistication have come to the same conclusion. Thank God for the Atom Bomb,” as Paul Fussell (Normandy veteran) once wrote. Or, as my former teacher Dr Knopfelmacher (Normandy veteran) remarked, “it knocked the b*llsh*t out of them”. And of course everyone forgets that Truman, unlike most of his second guessers, was a combat veteran himself. Experience makes a real difference.

There ought to be a law against war. But by the time war gets serious it gets tribal which means that high-minded moral reasoning is a subordinate consideration. The best we can hope for is to end it sooner rather than later. By that standard Truman’s dropping of the A-bomb passed the Machiavellian test.


Guido Nius 09.30.10 at 12:55 pm

Henri, well – as said, I do not think there is the kind of caricature utilitarianism you apparently need to believe in and I’m sure Kevorkian is somewhat of a nutter but having lived with people that expressly stated they would not want to go through Alzheimer, I think there is a good case for euthanasia. And if there is a good case for it, there might be a good case for using the bodies if they can do someone else some good. In any case it’s a better strategy than going all Christian on everybody with caricatures of doom & gloom and making everybody’s life miserable just to make sure that you can keep everybody living despite their own wishes.

(and I’m against the death penalty so the second Kevorkian quote is decidedly uncool)


Henri Vieuxtemps 09.30.10 at 1:18 pm

Hey, I don’t want to believe in the caricature, that’s what I’ve been saying all along. Utilitarianism is certainly a part of everybody’s decision making, to one degree or another.


Tim Wilkinson 09.30.10 at 1:35 pm

Jim Rose @203: I really did mean I’d have to look it up – but obviously not just by looking at Wikipedia!

I notice you seem stuck on the matter of whether the very same surrender would have happened at the very same time had everything been the same but for the detonation of those two bombs on those two cities. The actual ‘trolley’ question concerns what all the available options were and what their expected consequences were.

There is also the separate issue of the actual motives of the Truman mob – if they did as you claim take the best course when they opted, of all possibilities, to obliterate two cities, was that a foreseen consequence? An ‘intended’ one, in double-effect terms? (Double effect is perfectly intelligible, and definable in terms of counterfactual conditionals, when it comes to the issue of motive rather than of prospective bestness.)

On the assumption that you are as hostile to the washing of hands as you are to their wringing, you presumably have ready answers to other relevant questions: should casualties of {a: military personnel, b: civilians} who have the option of surrender be counted equally with those A-bombed without warning? Just to be clear, Japanese soldiers did actually surrender, didn’t they. Of those commonly included in death-or-glory stats, at least some, and probably quite a lot were killed after or during surrender, weren’t they. What about the fact that there was apparently increasing civil unrest among the supposedly implacable Japanese civilian population? Was this discontent reflected among the lower military ranks?

Should {c: conscript, d: regular} soldiers be counted equally with civilians (how much choice do we view {e: enlisted, f: commissioned} regular soldiers as having, and is that relevant)? Did the US have {g: the right, h: also the duty} to give {i: absolute, j: weighted} priority to {k: their own, l: allied} troops’ lives?

These are genuine questions – some answers could help you improve the case you have adopted.

A few methodological issues, too: in arguing your case, are you restricting yourself to information and inferences available to the US command at the time, or instead addressing the objective counterfactuals? (And are you doing better at counterfactual reasoning than those historians who tend for perfectly understandable reasons to look at what actually happened and add or remove small bits here and there, and treat that as sufficient to answer the questions they choose to set themselves? E.g. using the tractable question of causal influences on the Japanese surrender as a proxy for the question ‘how might casualties have been {m: minimised, n: minimised given a desire for swift Japanese surrender on US terms}?) In particular, are you giving the US administration knowledge of the details of the coup attempt? If there was such an attempt, are we to suppose that it would have been successful had A-bombs not been dropped on those particular cities when they were (which was later, I think?).

Re genie – by 1945 there was a genie, but was it really out of the bottle before Hiroshima? Was the significance of its unbottling lost on Stalin, and did it (on top of the Truman administration’s new anti-Soviet stance) have an effect on subsequent relations between the Soviet and American govts?

But anyway the key thing is that, assuming the bombs being dropped were necessary for a casualty-reducing surrender, how did they bring it about? Was it just by demonstrating that the US had the bomb and that it worked, for example? If so, did two cities really need to be immolated in quick succession? Did any cities?

There’s a nice cherry orchard for you.


Tim Wilkinson 09.30.10 at 1:54 pm

Jack Strocchi @208 speak good.

But I think Mill’s (or a Millian) account of indirect utilitarianism/conseq’ism is a bit more complex and subtle. There is the matter of stable dispositions, and of social insitutions of priase and blame as an informal alnague (or indeed extension) of official political institutions. Deep facts about humanity: we are not entirely good at, e.g., lying (some even blush – we may trust them more; some are very good, high-functioning psychopaths who unfortunately have an advantage in getting t othe top of politics..) we are not able to perform hard calculations (this also probably means we cannot be Bayesians, thus we get ‘paradoxes’ about knowledge like the Harman/Kripke intransigence problem), we need some predictability in other’s behaviour, etc. All these things are higher-order conseu’ist arguments for why there should be adherence to rules (pssst! for the most part). Which is not to say aetiology follows axiology here.

Also, going by the description, JQ’s position on the ticking time bomb stuff is very similar to (‘similar to’ being symmetric) me own position. The distinction between laws and individual ethics is deliberately elided by those who want to have state torturers sitting around twiddling their own thumbs while they wait for those of others to be delivered to the sterile dungeon when the stakes are high enough (which will occur somewhere short of the apolcalyptic scenario, one suspects.)


Tim Wilkinson 09.30.10 at 2:17 pm


Substance McGravitas 09.30.10 at 3:23 pm

Speculating on the the figurings of American strategists or the imminence of Nipponese surrender and $2.50 will get you a cup of coffee. No one doubts that “Truman had options”, but the best one was to end the war sooner rather than later. That is the option he chose for thos reasons and it was the right one, on utilitarian grounds alone.

It’s a weak utilitarian claim though: what is measured is potential losses, and the invasion plan mentioned was a plan among other options, not least was a moderation of the Potsdam language, which Truman indicated was possible. If the sticking point is, say, the preservation of the Emperor’s station then on utilitarian terms the dropping of the bomb is a horror.


geo 09.30.10 at 5:26 pm


Again: can you remember where the Chalmers Johnson quote in 195 is from?


Dave 10.01.10 at 1:02 pm

geo @216
It’s from a Chalmers review in the LRB:

Comments on this entry are closed.