Preliminary remarks: Recently we've been exploring the moral principles at work in abortion, euthanasia, and animal ethics.

And at the moment we're in somewhat of a dilemma between saying on the one hand that serious moral consideration is triggered by personhood, which entails the almost unrestricted permissibility of abortion, infanticide, and toddler-icide,

and saying on the other hand that serious moral consideration is triggered by sentience or being the conscious subject of a life, which entails the almost unrestricted impermissibility of killing or eating nonhuman animals not to mention late term fetuses, infants, and toddlers.

But even if we were entirely clear & correct about these issues, there would still be a further general question about the permissibility of killing innocent persons (and thus by implication, also the permissibility of killing any creature with less moral status than a person). It is widely believed that innocents can be permissibly killed in some circumstances but not in others. So what justifies the killing & what justifies morally discriminating between different types of circumstances? That's what the Trolley Problem is all about.

(1) Foot, "The Problem of Abortion and the Doctrine of the Double Effect"

--the basic argument-structure of the article: Foot criticizes & rejects the doctrine of the double effect (DDE) and then replaces DDE with a moral principle involving a distinction between negative duties and positive duties (NegPos)

--DDE: (1) direct intentions vs oblique intentions.

Someone's intention to do an act A is direct if and only if she wants to do A. By contrast, someone's intention to do act A is indirect or oblique if and only if she foresees that by doing B she will also end up doing A, and she wants to do B and thus directly intends B, but she may or may not want to do A.

(2) DDE = it is sometimes permissible to bring things about by oblique intention that are impermissible to directly intend
 

--Foot argues against DDE by appealing to various fictional cases that we evaluate by moral intuition. The cumulative effect of our intuitive judgments is that DDE is either outright false or acceptable only to the extent that it is consistent with a more basic principle = NegPos

--NegPos: (1) negative duties vs. positive duties

Negative duties are duties not to interfere with someone & in particular not to harm someone. By contrast, positive duties are duties to benefit someone & in particular to aid them.

(This should be compared with Foot's distinction in "Euthanasia" between liberty rights & claim rights. Negative duties are duties that correspond to liberty rights. Positive duties are duties that correspond to claim rights.)

(2) NegPos = in general, negative duties trump positive duties = in general, our obligations to refrain from harming people are fundamentally stronger than our obligations to give aid to people = in general, it is worse to harm people than it is to fail to give aid to people = in general, in cases in which there's a choice between not harming people and giving aid to people, you should always go for not harming people even at the risk of failing to give aid to people = in general, you're always required to avoid harming people but not always required to aid people.

--other distinctions considered by Foot:
        doing vs. allowing
        commission vs. omission
        doing vs. refraining
        active killing vs. passive killing

--cases: abortion cases:
        mother's life vs. fetus's life 1 (both will die if nothing done)
        mother's life vs. fetus's life 2 (fetus will live if nothing done)
        fat man in mouth of cave (head in vs. head out)
        wicked merchants
        wicked grave diggers
        judge in riot
        pilot in crashing plane
        trolley driver: kill 1 to save 5
        drug: 1 vs. 5
        single patient vs. multiple crash victims
        killing people to use their bodies for cancer research
        serum: kill 1 to save 5
        bad man 1: torture 1 or he'll torture 5
        bad man 2: kill 1 or he'll kill 2
        poisoning vs. allowing to die by poison
        allowing people to die of starvation in Africa
        lethal gas
        cannibal sailors

--now compare & contrast the two cases in boldface => Foot's Trolley Problem is this: why is it permissible for the trolley driver to kill 1 innocent in order to save 5 other innocents, yet impermissible for the doctor to kill 1 innocent in order to save 5 other innocents?

Foot's answer is: NegPos. The trolley driver has a choice between harming 1 or harming 5, so it's better to harm 1, but the doctor kills 1 & thus harms that 1 in order to aid 5. So by NegPos, trolley driver is permissible but serum is impermissible.

(2) Thomson, "Killing, Letting Die, and the Trolley Problem."

--the basic structure of Thomson's article is analogous to Foot's: Thomson criticizes & rejects the killing vs. letting die principle (KLD), and replaces it with a distinction between doing something to a threat vs. doing something to a person (Threat/Person)

--KLD: Killing is always morally worse than letting die.

--as with Foot's argument against DDE, Thomson argues against KLD by by appealing to various fictional cases that we evaluate by moral intuition. Here the cumulative effect of our intuitive judgments is that KLD is either outright false or acceptable only to the extent that it is consistent with a more basic principle = Threat/Person.

Along the way, Thomson also uses the same method to reject NegPos: Frank the trolley passenger is clearly permissible, but violates a negative duty to avoid harming people

--but most importantly, Thomson isolates Foot's Trolley Problem as capturing the essence of the moral issues involved,

and she also introduces an even sharper version of the problem in the contrast between Frank the trolley passenger and George on the bridge =

why is Frank's act of killing 1 innocent to save 5 other innocents permissible, while George's act of killing 1 innocent in order to save 5 other innocents is impermissible? Let's call this "Thomson's Trolley Problem."

--Threat/Person:
Assuming other things are equal, it is permissible to kill 1 innocent person in order to save other 5 innocents if and only if there's a threat to the 5 that can be redirected to the 1 by doing something to the threat but not doing something to a person in the redirecting act.

(In a later article, Thomson unpacks Threat/Person a little more by specifying that doing something to the threat = deflecting an existing threat as opposed to introducing a new threat, and by specifying that not doing something to a person in the redirecting act = violating no one's rights in the deflecting act as opposed to violating a liberty right or a claim right as a part of the deflecting act.)

--cases:
        Alfred vs. Bert
        Alfrieda vs. Bertha
        Charles the surgeon
        David the surgeon
        Edward the trolley driver
        Frank the trolley passenger
        George on the bridge
        Harry the president
        Irving the president
        health pebble cases
        variants on Edward the trolley driver & Frank the trolley passenger
        avalanche
        Donald the surgeon

--worries: while I agree with Thomson that Foot is wrong about NegPos, I think that there are counterexamples to Threat/Person

E.g., suppose that in order to get to the steering wheel in time to turn the trolley onto the spur so that he can kill the 1 in order to save 5, frank the trolley passenger has to push another passenger out of the way (say, thereby breaking the other passenger's arm). Let's call this Frank the arm-breaker. Now I think that if Frank the trolley passenger is permissible, then so is Frank the arm-breaker. But Frank definitely violates someone's liberty rights. So Threat/Person must be false.
 

            but if Threat/Person is false, then what's the solution to Thomson's Trolley Problem?

Now I'm inclined to think that the difference between Frank the trolley passenger and George on the bridge is that in the latter the fat man is an uninvolved person who's brought into a threat-situation without his consent, whereas the 1 on the track is initially involved in the threat-situation, and that more generally the crucial moral factor is the brute factual difference between being initially involved or uninvolved in the moral situation: our obligations to people are different depending on where they just happen to be placed in space or time relative to an ongoing moral situation.

But if I'm right, then in effect I'm committed to some sort of "principle of moral inertia," which Thomson thinks is clearly wrong (p. 209) => how can some purely factual & contingent feature of a moral situation ever generate a moral obligation? => wouldn't this entail deriving an ought from an is?