One act, two effects: willing the good, accepting the foreseen harm

The doctrine of double effect

Thomas Aquinas 1225 · 1274Faith and reason cannot contradict each other.

What it claims

An act with a good and a bad effect is permitted if the act is good or neutral, if the bad effect is aimed at neither as an end nor as a means to the good, and if there is a proportionate reason to accept it.

Why

An act is judged by what is intended, not by what is foreseen but not aimed at: an unintended bad effect can be allowed.

The same idea through history

Thomas Aquinas1271Around 1271, on self-defence: one act can have two effects, one intended, the other beside the intention; the act is judged by what is intended, provided it stays proportionate to its end.

Jean-Pierre Gury1850Gury's handbook of moral theology turns it into a general principle with four conditions: an honest end, a good or neutral act, a good effect that does not come through the bad one, a grave and proportionate reason. This is the form taught ever since.

Philippa Foot1967Foot presents the doctrine in its strongest form (to intend strictly is to aim at something as an end or as a means), applies it to the trolley, then concludes that it plays only a secondary role.

In short

"One act, two effects: willing the good, accepting the foreseen harm" is an argument by Thomas Aquinas (Thomas Aquinas, Summa theologiae, II-II, q. 64, a. 7 (c. 1271); Gury (1850); Foot (1967), 1271) on the question "What makes something right?" (answer: duty). Its claim: An act with a good and a bad effect is permitted if the act is good or neutral, if the bad effect is aimed at neither as an end nor as a means to the good, and if there is a proportionate reason to accept it. According to AskPhilosophy's logic check (Z3 solver), the argument is valid only if one grants a premise the author leaves unstated (P2). AskPhilosophy presents 4 objections, 2 of them with a documented reply, and the original text behind each premise, checked word for word.

Thomas Aquinas, Summa theologiae, II-II, q. 64, a. 7 (c. 1271); Gury (1850); Foot (1967)

Valid with an implicit premise

A man attacks you to kill you. You defend yourself, and your blow kills him. Are you a murderer? Thomas Aquinas answers no, if you wanted to save your life and not to kill: the attacker's death was not your aim, only a foreseen result of your defence. But he adds a limit: if you strike harder than necessary, the defence becomes unjust. Later, moralists turned this answer into a general rule, the "doctrine of double effect": one may never intend evil, either as an end or as a means, but one may sometimes accept an evil one foresees without intending it. It is used in medicine (a painkiller that shortens life), in war (bombing a factory knowing civilians will die) and in the trolley problem: at the switch, the workman's death is a foreseen effect; on the footbridge, pushing the fat man makes his death the means.

click a premise to reject it

If you accept every premise, logic forces you to accept the conclusion. So the debate is about the premises.

Checked by a logic solver: the premises cannot all be true while the conclusion is false.
Without the implicit premise: Without P2, the computer finds a world where the bad effect is neither the end nor the means, but is still intended: the act is then no longer good in itself, and nothing guarantees it is permitted. That is exactly the "closeness" problem: does what is stuck very close to what we aim at, without strictly being the means, count as intended?

Valid with P2. Thomas only says "intended" or "beside the intention". The modern conditions speak of end and means: so one must add that to intend is to aim at as an end or as a means (the definition Foot gives). The debate is mostly about P2 (where does what we intend stop? the potholer, the loop) and P1 (does intention really change what is permitted? Foot's gas case).

Analogy. A doctor gives morphine to a patient in great pain, knowing that at a high dose it may hasten death. If he wants to relieve the pain, most judge him within his rights. If he gives the same dose so that the patient dies, it is something else, even though the act is the same. The doctrine says this difference in intention matters.

Objections

Each objection is a conversation: click to see who replies and where the debate stands.

H. L. A. Hart and Philippa Foot ("closeness") 1967 · attacks premise 2

Potholers are trapped in a cave filling with water: the fattest of them is stuck in the exit. They have dynamite. Could they say: "We did not want to kill him, only to blast him out of the exit; his death is only a foreseen effect"? That is ridiculous, and the doctrine's defenders reject it. But then where is the line? If whatever is "very close" to what we aim at counts as intended, we need a criterion of closeness, and nobody has given one. Hart also noted that in craniotomy (crushing an unborn child's skull to save the mother), the child's death is not, strictly speaking, the means.

Thomas d'Aquin replies

Warren Quinn 1989 · Quinn's reply: what is done to the victim

The doctrine must be reformulated. What matters is not whether one intends the death itself, but whether one deliberately involves the victim in one's plan, acting on him to achieve one's end. The potholers mean to blow up the fat man's body: they act on him, which is forbidden, even if his death is not their aim. At the switch, the workman plays no part in the plan.

The debate goes on (our summary) Quinn's version avoids the potholer, but it is no longer quite Thomas's doctrine: it is no longer about what one intends in one's head, but about how one treats someone. Critics say blurry cases remain; defenders say every moral rule has borderline cases without being false for that.

Philippa Foot (the gas) 1967 · attacks premise 1

Five patients in a hospital can be saved by making a gas, but making it sends lethal fumes into the room of a sixth patient who cannot be moved. His death serves no purpose: it is a pure side effect, and with five against one, the proportion is there. So the doctrine should permit it. Yet we judge it forbidden, as if we were killing the patient. For Foot, what explains our judgements is not the difference between intending and foreseeing, but the one between harming and not helping: the duty not to harm is stricter than the duty to help.

no reply

No reply possible: objection raised in 1967, after the author's death (1274).

The debate goes on (our summary) Defenders reply that the doctrine gives necessary conditions, not always sufficient ones: an unintended bad effect still has to be justified, and proportion is not just counting lives (the patient's rights count too). But then the doctrine decides less than expected, which is exactly Foot's conclusion: intention plays "only a quite subsidiary role". Thomson (1985) also judges the gas case plainly forbidden.

Judith Jarvis Thomson (the loop) 1985 · attacks premise 2

Change the side track: it loops back towards the five. The trolley will stop only if it hits the lone workman, whose body will block it. This time the workman's death really is the means of saving the five: the doctrine should forbid turning, as with the fat man. Yet a little extra bit of track cannot make much moral difference, and it seems one may turn. If so, the doctrine gives the wrong answer here.

Thomas d'Aquin replies

Frances Kamm 2007 · Kamm's reply: acting "because of", not "in order to"

In the loop, one does not turn in order that the workman be hit: one turns to move the trolley away from the five, and does so because the workman, by blocking it, will prevent its return. Acting while counting on an effect is not aiming at it. The doctrine, made precise in this way (Kamm calls it "triple effect"), can accept the loop and reject the fat man.

Awad et al. (Moral Machine) 2020 · The data

Across 70,000 people in 42 countries, the loop is always judged in between: less acceptable than the switch, more than the footbridge. The authors note that many find it hard to say whether, in the loop, the workman's death is intended or merely foreseen.

The debate goes on (our summary) The loop remains one of the most debated cases. For Thomson, it shows that neither the doctrine of double effect nor the idea of "using someone" explains the trolley problem; she looks for the answer in rights instead. For defenders, it only forces us to say more precisely what "intending" means.

Joshua Greene 2010 · attacks premise 1

The difference between means and side effect, which we feel so strongly, could be a by-product of how our brain represents actions: the emotional alarm system that reacts to violent acts "sees" means and not side effects. If so, this distinction describes our psychology, not what is really permitted, and we should not take refuge in it.

no reply

No reply possible: objection raised in 2010, after the author's death (1274).

The debate goes on (our summary) Defenders reply as Berker did to Greene on the trolley: explaining where an intuition comes from does not show it is false. The distinction between intending and foreseeing also exists in law (murder versus manslaughter), and it has been defended with reasons, not just felt. Greene announced the details of his explanation for a forthcoming book (Moral Tribes, 2013).