Allowing evil that good may come of it?

Huge amount to respond to here Alex

I like how you have characterised the dilemma and avoided the third option, which to be frank seems to be a bit of a cop out.

I discussed this at great length in my MA thesis a few years back: God the Utilitarian? The ethics of theodicy

I think there are several difficulties. Firstly, it is hard to see whether the doing/allowing distinction can do much work when we are talking about an omnipotent being, if the Rachels/utilitarian counter examples show us anything it is when intervention is almost cost free then there doesn't seem to be a strong distinction between doing and allowing. And again, being the creator makes it difficult, especially with predictability, to see how God could just be allowing things to happen. Furthermore, if you take the divine underwriter approach, so God doesn't just set the ball rolling, but also underwrites its rolling it seems hard to see how only allowing is happening.

I'll follow up with further difficulties for the position later.

Alexander,

Interesting post! There's a smallish step in the argument that I wonder about. You write, "The worry is that if we allow God to be moved by this consideration, then God will be intending Jones to develop patience, and hence will be intending the means to that, namely the destruction of Jones' legs, and hence God will be willing an evil." Do you think this step is just analytic? In light of the epistemological discussion about transmission (or closure) principles, I wonder if some similar considerations apply to transmission principles for intention.

My understanding is that since Aquinas this sort of difficulty has usually been resolved by something like a double effect principle. That doesn't seem to fit into your structure, either, so perhaps that's a fifth horn of a pentalemma (is that the right term?).

Jeremy:

Double Effect seems to be the way to go, but that is what generates most of my concerns, because here the evil appears to be a means to the good, and this is forbidden by Double Effect.

Ted:

Yes, I am assuming that if A is one's means to B, and B is intended, then A is intended.

David:

I think the difference between doing and allowing is metaphysical: it is a question of the metaphysics of the will--action is something that the will's causality is positively behind or something like that. So, yes, I think the distinction is significant even in costless situations. The reason it is wrong to actively do evil is that then one's will is aligned contrary to its loving nature or something like that.

I intend to go to my son's room and my means to go to my son's room is to walk through the living room. But I don't intend to walk through the living room.

I think you need to supplement the principle with "intentional mean". But then why think every mean is an intentional mean?

Ted, you say,

But I don't intend to walk through the living room.

Suppose you walk over the newly cleaned carpet in the living room. Can you plausibly plead that your walking over the carpet in the living room was unintentional? I doubt it. Unless you did not see the carpet or did not recognize it as the living room carpet. But we can stipulate those away. But then you intentionally walked over the carpet in the living room and did not intend to walk through the living room? Maybe you'd want to say that you intended to walk through the living room but you did not intend it as a means to getting to the child's room. Still it is difficult to see that.
A surgeon might cause pain intentionally--he prepares the scalpel, applies it to your abdomen, makes a deep vertical incision and knows that this action is painful to you (or soon will be). He cannot reasonably plead, "sorry, that was unintentional". But he can plead that he did not intend to cause pain as a means to helping the patient. The pain does not help at all. In the living room case, you intend to walk through the living room, and it is hard to see that you do not also intend it as a means.

Mike,

That sounds right, but note that you've strenghtened the principle to make it more immune to my counterexample. It's probably the case that there's some counterexample-free principle that does the work it needs to do in Alexander's argument. I just wanted to push against it and see where that led. Here's another case to think about. Can I intend that 2+2=4? If not, then what's the difference between that and intending E through M where M *must* occur for E to occur?

Ted,

That's an interesting question. Since 'intends to' generates opacity, wouldn't you say that I might both intend the impossible and intend the necessary? I intend to make 2 the smallest prime for me, believing that it's a matter of individual taste. Or I intend to shake Ortcutt's hand and never to shake the hand of the man in the brown hat. In the former case I intend the necessary and in the latter I intend the impossible. These look like things I can intend.
It is true, I think, that I cannot intend to do what I believe I can't do (though, I like examples to the contrary such as 'Smith intended to dead-lift 400lbs, though he believed he couldn't do it. Maybe 'trying to X' and 'intending to X' pull apart in such cases). I'm not sure whether it is true that I can intend to do what I believe I must do. If I leap from the cliff, is my falling intended? Seems like it, though of course it must occur, if I leap. Suppose you leap onto someone in the pool: "I did intend to leap, but I didn't intend to fall, so my falling on you was unintentional". If I were the one in the pool, I wouldn't buy it. Still, these might not be the sorts of cases you had in mind.

Mike:

I think it's correct to say that the surgeon is not intending to hurt the patient. "Unintentionally" is probably a bit stronger than "not intending to". We may in the end need to distinguish things like: what one intends, what one does intentionally, what one does not unintentionally. All this is hard to do. My official account of Double Effect in terms of accomplishment tries to help with these issues.

I also think we need to distinguish what is typically psychologically necessary from what is necessary as a matter of the structure of action. Thus, psychologically, it's pretty much impossible to leap without intending to fall, in most circumstances.

Here's one way to generate double effect style claims. Thesis: The will and the intellect are distinct in such a way that one's acquiring a new belief does not necessitate a change of will. Intention is a matter of will rather than belief.

Suppose now that I falsely believe I am in zero gravity, so I believe the leap will not be followed by a fall. I intend to leap, because I care only about the exercise of my legs--the exercise of my legs is all I care about. I don't care whether I live or die. I just want to stretch my legs. (My will is obviously screwed up, but nonetheless the situation seems possible.) I now learn that there is gravity, and conclude that I will fall after leaping. Merely acquiring a belief does not change my will automatically. Thus it is possible that I will still intend only the leap and not the fall. (Of course I will still be guilty of endangering my life, since I will be neglecting the ceteris paribus duty to protect my life. I will also be guilty if there is someone at the bottom on whom I will land.) The additional information about the presence of gravity may simply be irrelevant to me volitionally.

I think it's correct to say that the surgeon is not intending to hurt the patient.

I didn't say that he intended to hurt the patient, but there is a good chance that he does. He just does not intend it as a goal or as a means to a goal. When the patient says "ouch, that hurts. Did you intend to do that?" I think his answer has to be "yes, but I certainly did not intend it as a means or as a goal". He could not say "Oops, I didn't intend to do that. Sorry."

But what I said was that he caused pain intentionally. Maybe a distinction can be made between, X intended to do A and X did A intentionally. I have no doubt that the surgeon caused pain intentionally. It certainly was not unintentional. It would be surprising to learn that there are things I do intentionally but don't intend to do.

Dear Mike:

Well, first there is a category difference. Only actions are done intentionally, but other kinds of events than actions can be intended.

But the more significant difference is that I think in ordinary language "he intended p", or "he aimed to bring it about that p" (that may be a better locution), is more intensional than "he intentionally did A". I suspect that "he intentionally did A" holds if (though perhaps not only if) the agent did something that he knows to fall under A.

Best wishes,
Alex

I'll still not convinced that a creator can just allow anything, but I take your point about the metaphysical act of willing.

Here is a general problem for these sorts of positions, if it is good for God to allow evil, without intending it, then should we not likewise allow evil as long as we don't intend it? After all we know (due to God allowing it) that it is ultimately likely to bring about a greater good.

On the character development style theodicies, the difficulty I have always had with them is that many of the things we regard as virtues seem only to be virtues because the world has characteristics that make those things good. Take stoicism, this is good only because there is suffering. So allowing suffering to generate stoicism doesn't seem to work.

Sorry, I maybe going a bit off point, but it seems to me that when St. Paul forbids us to do evil that good may come out of it suggests a certain deontological strain to ethics.

Fyodor Dostoyevsky has already straight away cut to the chase by asking the question of could there possibly be any future good (such as a future utopia of people who love God out of their free will) which would justify the the horrendous sufferings of a single child?

If one take a deontological structure to ethics, then it seems that no good consequences ever justifies the horrendous suffering of a child, and the whole question of whether God intentionally permitted evil or permitted it without intention is irrelevant to theodicy.

If God knew that permitting evil would result in the cruel suffering of a single child, and if, for some inexplicable metaphysical reasons, God cannot achieve His utopia except by the permission of evil (this is to side-step the question of whether God needed to give free choices or "grow" spiritual people by inflicting them with evils), then, deontologically speaking, God should not create anything at all since to create would result in horrendous suffering of a child which is never justified or which no good ever will make right.

It seems that if one is committed to (3), that ethics is not essentially different for God and man, then it seems that this theodicy entails that everyone should be consequentialists, which has never really sat well with theists in general.

I think I would prefer to reject (3) rather than to enter into the murkier waters of trying to justify evils with good consequences.

David:

I think the argument that we should allow evils fails. The question is only urgent for evils that it is within our power to prevent. But for such evils, one of God's defeaters for the claim that he should prevent the evil may be that in preventing it he would also make it harder for us to fulfill our task of preventing evils of that sort--preventing these evils is our task. Now if we don't prevent the evil, then the good of our having the opportunity for having prevented it is still there. But it would have been even better if we not just had the opportunity but also took the opportunity.

Nor am I claiming that every evil can be gotten around in this way. First of all, only evils that are allowed and not directly caused by God need apply for this kind of a story. (E.g., it will not justify God lying.) Second, this is only going to apply to evils where the defeater of defeater structure of action applies. So these must be evils that not only result in greater goods but that are themselves the result of significant goods.

Finally, one still has the highly non-trivial task of standard theodicy, namely of ensuring that the goods in question are of sufficient magnitude.

Dominic:

If I had the choice to prevent the horrible sufferings of one child or to stop a device about to release a nerve gas that will cause millions to die a painless death, and I did not have time for both, I think it would be at least permissible for me to opt for the latter. So it is false that nothing can justify permitting the horrible suffering of one child.

This is not really an answer to what you say. For you do not say that nothing can justify permitting the horrible suffering of one child. Rather, you say that no good can justify it. In my example, an evil is involved. Nonetheless, my example may still tell against the principle because the prevention of an evil is a good, so the prevention of the deaths of billions does seem to be a counterexample to the principle. (That this particularly case can't come up for an omnipotent being does not affect the fact that the moral principle is shown false by the case.)

Moreover, it is far from obvious to me that it is wrong to prevent the horrible suffering of a child for the sake of an infinite good to the child. Suppose that the child is very poor, and that by tarrying for one minute to the rescue from the horrible sufferings, I could ensure that the child's college education be paid for. Tarrying might be acceptable. Now suppose that one multiplies the good by infinity and the length of time tarried by a finite amount (say, a hundred or a thousand). It seems like tarrying still might be acceptable.

Alex, you write,

But the more significant difference is that I think in ordinary language "he intended p", or "he aimed to bring it about that p" (that may be a better locution), is more intensional than "he intentionally did A".

I suppose investigating ordinary language is the way to make these distinctions. I agree they're hard to make. The only discussion that came to mind was Austin's 'Three Ways of Spilling Ink'. I'm not sure what you mean by these terms being of different categories; I agree of course that they're different parts of speech, one's a verb the other's an adverb. But that does not mean there aren't important logical relations between them, and that is what I was (more or less) asking about. I was looking for a case in which 'I intentionally did X' does not entail 'I intended to do X'. For instance, 'I intentionally caused the pain' seems to entail that I intended to cause the pain. I don't think it entails that I intended to cause pain as a goal or as a means. I don't think it entails (not to make things more complicated) that it was my purpose to cause pain.
In any case, both the verb and the adverb seem to me intensional. I intend things under certain descriptions, and I intensionally do things under descriptions. I guess I'm not sure what you mean by one being 'more intensional' than the other.

I can think of two ways a deontologist could get around this problem by denying a particularly Kantian deontology. One way is to say that deontological claims but limits on what we can do but not in an absolute way. So the mere fact that outcome A is better than outcome B won't be enough to justify doing something to attain A that violates a deontological principle. It would have to take a particularly serious outcome difference to justify it. Deontology remains as the denial that any old consequences difference can justify doing what would otherwise be wrong. The deontological principles are more than rules of thumb. They are almost always going to apply. But not always, and God should know when they do and don't apply. I'm not sure if this fits any of the three options above.

Another way that might amount to the same thing in the end is to have a consequentialist foundation of deontology at the level of application. What makes it wrong for me to kill is that killing almost always leads to bad results. The ethical theory that will produce the best results is deontology (as followed by humans, anyway). Thus we should be deontologists. But the foundation is consequentialist, and the reasons for humans to be deontological go back to the same reasons an omniscient being would be a consequentialist (or at least of the modified deontology of the first paragraph of this comment, which is why this might end up looking like that in the end but just applying to God). So maybe that's a reason someone might like option 3.

Let me re-summarized the structure of the proposed non-consequentialist solution.

God intends to produce a good G (say, orderly laws). God's producing G, together with actualizing other intended goods, causes the existence of an evil E which is not a means to G. In the easy cases of deontological theodicy, one can use double effect directly here: God intends G and E is an unintended side-effect. But there may be times when E is so horrendous that it is out of proportion with G, and the proportionality condition in the principle of double effect will not be met (that condition says that the bad effect must be proportionate to the good).

In those cases, more work is needed. And my proposal is that sometimes this additional work will have the following structure. The badness of E provides a defeater for God's reason for producing G. But there is a second good, H, to which E is a means. Now God cannot intend H, since then he would intend the means E. But the existence of H can be a defeater for the defeater that comes from the badness of E. I think this works best when G2 happens to the same person as E does.

I think we can with some imagination come up with non-divine cases like that. The only ones I can think of are a bit outlandish. Here's a sci-fi one. I am the captain of a spaceship. Jim, a passenger who is a friend of mine insists on taking a side-trip, while we are re-fueling, to a dangerous planet. I warn him against it, but he says that he really wants to do it. Moreover, he begs me that if he is in danger, I shouldn't intervene--he wants to prove himself, or something like that. So off he goes. Jim instantly falls into the hands of the evil Dr. Zmog, who is now torturing him horribly. However, Dr. Zmog has an interesting custom that few people but I know about. After torturing people, Dr. Zmog bestows on them great intellectual gifts.

I am not able to communicate with Jim, but I can, at significant expense to the company I represent (but within my discretion--I am permitted to undertake this expense) beam him back.

Now I reflect:
1. I have two reasons not to beam Jim back: (a) Jim's autonomous request not to be rescued, and (b) the expense of beaming him back.
2. These reasons, however, have a defeater: the particularly horrible nature of the torture Jim is undergoing. I could argue that I will leave Jim be, and not intend his sufferings. But proportionality is not met: the evil is disproportionate to these two reasons.
3. But there is a defeater to the defeater: with the enhanced intellectual powers, Jim will be able to attain his cherished life goals (let's suppose he is a scientist). It would be wrong for me to intend for him to be tortured that he might gain such intellectual powers. But this further consequence of the torture, defeats the defeater that the torture represents.
4. It can be argued that in this case, in leaving Jim in Dr. Zmog's hands, I am intending only to keep Jim autonomous and, to a lesser extent, to keep down the expense to the company I represent. The fact that the torture has a silver lining lets me screen the horribleness of the torture from my weighing, maybe.

Mike:

Ordinary language is mainly a way of softening up someone here to notice a distinction. I think there is a distinction to be made. "Intention" may not be the best term for it always "What is aimed at (ultimately or mediately)" might be better. Certainly, "What is aimed at" works really nicely literally in the standard bombing cases of double effect (Jones the bomb aimer, whose bombs fall on both the enemy HQ and the civilians, is aiming at (bombing) the HQ and not the civilians).

I think one way to make the distinction more rigorously is to have a particular metaphysics of the will. For instance, one might hold that in a choice the will gets "set" (in some way to be explicated further) on a particular description of a state of affairs, and its being set on that description causes the person to engage in a striving in the direction of the existence of something filling that description. The thing described might be an end or a means. The agent may or may not be conscious of the description. (In this way, the description is very much like a Kantian maxim. We are quite fallible in telling which maxim we acted on, but what we are responsible for depends crucially on that question.)

By the way, my best attempt at an account of Double Effect is at AlexanderPruss.com/papers/PlansAndTheirAccomplishment.html

Jeremy:

An "in extremes, it's OK to do really bad things" deontology fits poorly with the Christian ethical tradition, at least in the West. It also fits poorly with the Jewish tradition which holds that there are some (very few: idolatry, adultery and murder are the three that I remember, but there may be one or two more) things that may not be done no matter what.

I am concerned, too, about the fact that, barring vagueness, there would then have to be a least integer N such that, when one knows nothing about the N+1 persons involved, it is permissible to kill an innocent person to save N lives. What would that N be? 17? 16? It seems an odd fact of the matter as to what N is.

The option that consequentialism is God's ethics is attractive. I've heard this view ascribed to Leibniz: God creates the best possible world, and then commands us to keep to those laws commanding us to keep which will make for the possible world; our duty, then, is not to maximize the good, but to keep to these laws. (I've seen texts of Leibniz that do not fit with this view, and I haven't seen texts that do fit with it, and I can't remember where I heard the attribution.)

I think this view fits nicely with a divine command theory of human ethics. That said, I do not find it attractive. For one, it would make God a maximizer of the good, and that would induce decision paralysis, since in fact it seems there is no best world.

The view I have in mind is not accurately describes as saying "in extremes, it's OK to do really bad things". The point is that they're not really bad in these cases. The idea is that moral principles have a threshold for consequences above which the constraint against doing the action doesn't apply. It's not that you violate the constraint but that it's ok to do so. It's that it simply doesn't apply. In a life-or-death circumstance, it would be wrong for me to keep my promise to meet someone for dinner. It would be wrong to worry about damaging my neighbor's lawn. Perhaps in both cases I owe something to the people in question, but in both cases I don't just do nothing wrong. I do the right thing. Yet I think you can say all that without being a consequentialist about when it's ok to break promises or damage your neighbor's lawn.

I should also say that there might be some things that are so bad that there is no possible situation bad enough to overcome the constraint.

Jeremy:

Right. A more fair formulation is that under extreme circumstances it is OK to intentionally cause intrinsically very bad states of affairs. (The death of the innocent person killed to save a billion is still a very bad state of affairs.)

I think promises to do something morally wrong are not binding, insofar as they are promises to do something wrong. To to come for dinner even if this would prevent me from saving a life would be immoral. Hence the promise is not binding in this respect at all.

Hmm, it seems that the solution proposed by Jeremy is a form of rule consequentialism, where deontological rules are justified by good consequences, and analysis of moral decision is to be made at the level of rules rather than at individual actions.

To respond to Alexander, I would not say that allowing the horrible suffering of the child is morally permissable at all in the light of saving million from being gased. I find both acts equally repugnant, and if I were forced into such a choice and if God is someone who is even remotely like a human moral agent, I would curse God for having placed me in such a terrible moral conflict in the first place.

Bernard Williams have criticised such forms of moral thinking in "Ethical Consistency", the form of moral thinking that demands that moral obligations cannot "really" conflict and that if they do, one of them has to be "merely" a prima facie obligation. Such thinking does not do justice to the phenomenon of ineliminable agent-regret.

Thus, I see in the example of the suffering child versus saving millions a fundamental conflict of morals which both are equally morally repugnant.

To use an example given by D.Z. Phillips, he speaks of a true story of Sophie Zawistowska, a Polish survivor of the Auschwitz who attempted to steal a radio from a German commander to aid the resistance. Caught, they told her that out of her two children, they would kill one of them, but which is up to her to choose. If she did not choose, they would kill both of them, to which Sophie cried, "Ich kann nicht wahlen!" or "I cannot choose!" But in order not to loose both of them, she choose to let her son free and they killed the daughter. Sophie's "choice" stayed with her, and after she was liberated, her choice realized itself fully on her and she committed suicide two years after.

Of course, under a gross consequentialist analysis, it is better that one child survives rather than have both child killed, but the point of the story is that *both* options is equally repugnant. To not choose would result in the death of both, to choose will be to determine the death of one, the latter is no more "permissable" or "less morally bad" than the former, simply by virtue of a greater numerical good! To tell her that it was irrational of her according to some cool consequentialist calculus to regret having to choose is callous and would not to do justice to the what Williams calls the "inelimnable agent-regret". And who can blame her for her suicide?

From a third-person point of view, we might consider Sophie an object of pity and compassion, she cannot get over the moral significance of handing over her daughter, she was involved in a moral tragedy where, whatever she did, would involve evil. She does not think that the handing over of her daughter to be excused in the light of the total situation.

Does God think in the same way? What's God view on what He has done? Does He think that the Holocaust can be excused in the light of the greater good that made it necessary? Is creation a moral tragedy in which God is necessarily involved in evil? Or does God callously simply shrugged off the moral significance of the evil simply because it was justified by a greater good? Was God tormented and conflicted within Himself at the moral tragedy which He is involved in?

To give one more example, Ursula Le Guin, told the story of "The Ones Who Walk Away from Omelas", spoke of a utopian land, Omelas, where the fields without fail yield their abundance, the people were all joyous, where there was no suffering or crime, and where everyone was blissfully happy.

However, this land holds a terrible secret. As Ursula describes it,

"In a basement under one of the beautiful public buildings of Omelas, or perhaps in the cellar of one of its spacious private homes, there is a room. It has one locked door, and no window. A little light seeps in dustily between cracks in the boards, secondhand from a cobwebbed window somewhere across the cellar. In one corner of the little room a couple of mops, with stiff, clotted, foul-smelling heads, stand near a rusty bucket. The floor is dirt, a little damp to the touch, as cellar dirt usually is. The room is about three paces long and two wide: a mere broom closet or disused tool room. In the room a child is sitting. It could be a boy or a girl. It looks about six, but actually is nearly ten. It is feeble-minded. Perhaps it was born defective, or perhaps it has become imbecile through fear, malnutrition, and neglect. It picks its nose and occasionally fumbles vaguely with its toes or genitals, as it sits hunched in the corner farthest from the bucket and the two mops. It is afraid of the mops. It finds them horrible. It shuts its eyes, but it knows the mops are still standing there; and the door is locked; and nobody will come. The door is always locked; and nobody ever comes, except that sometimes-the child has no understanding of time or interval-sometimes the door rattles terribly and opens, and a person, or several people, are there. One of them may come in and kick the child to make it stand up. The others never come close, but peer in at it with frightened, disgusted eyes. The food bowl and the water jug are hastily filled, the door is locked; the eyes disappear. The people at the door never say anything, but the child, who has not always lived in the tool room, and can remember sunlight and its mother's voice, sometimes speaks. "I will be good," it says. "Please let me out. I will be good!" They never answer. The child used to scream for help at night, and cry a good deal, but now it only makes a kind of whining, "eh-haa, eh-haa", and it speaks less and less often. It is so thin there are no calves to its legs, its belly protrudes; it lives on a half-bowl of corn meal and grease a day. It is naked. Its buttocks and things are a mass of festered sores, as it sits in its own excrement continually."

All the people of Omelas knows its there. All of them knew that their happiness, the beauty of their city, the tenderness of their friendships, the health of their children, the wisdom of their scholars, the skill of their makers, the abundance of their harvest and the kindly weathers of their skies, depend wholly on this child's abominable misery.

The terms were very strict and clear. If the tiniest comfort or improvement of the child is done, if she is brought into the daylight, cleaned and fed and comforted, then in that day and hour all the prosperity and beauty and delight of Omelas would wither and be destroyed.

This would be explained to all the children when they are between eight or twelve. And when they see the child, no matter how well the matter has been explained to them, these young spectators are always shocked and sickened at the sight. They feel disgust, which they had thought themselves superior to. They feel anger, outrage, impotence, despite all the explanations. They would like to do something for the child. But there is nothing they can do.

To exchange the grace and goodness of all the lives in Omelas for that single improvement, to throw away the happiness of the thousands, is certainly to bring guilt into the walls of the city.

Many of the children would go home in tearless rage when they have seen the child. But as time goes by, they start to justify the torment of the child to themselves. She was too degraded to be able to get much improvement from the outside world anyway or something like that.

But there are others, who after having seen the child, doesn't return home, they walk out of the beautiful gates of Omelas, they walk ahead in the darkness, the place they are going is unimaginable, but they seem to know where they are going, the ones who walk away from Omelas.

Is the Kingdom of Heaven like Omelas? Are the saints there like the citizens of Omelas? Is the bliss of etenity purchased and build on the sufferings of millions of children? Do they justify and comfort themselves with consequentialist explanations? I think then that we should follow Ivan Karamazov, and return our tickets.

As to the tarrying of the child suffering and college example, let put it in another way. Suppose the child was told that all the child had to endure was one minute of being raped brutally and eternal bliss would be hers, would that justify the rape? Would any sane parent subject their child to such a thing to purchase eternal bliss? It sounds more like a deal from the devil! After all what's a "finite" amount of pain and suffering compared to the infinite of bliss right?

I think that we should stop theoreticising God as an economist and be true to the moral facts rather than try to rationalize all our ethical experiences together into a cold consequentialist framework.

I think that there does exist in reality, fundamental moral conflicts, which cannot be systematized away by any morality system.

To link it back to the post, no, there is no way in which an evil can be justified by a good.

Dominic:

I do think you're sliding between different cases. A case where the suffering of one causes bliss for many others sounds much more like a case where an evil is intended as a means to the good.

The case of Sophie is actually ambiguously formulated. I think there are two variants. Variant 1: Sophie must say: "I want this one to die." Variant 2: Sophie is permitted to say: "I want this one to live." Consequentialistically, the two situations are the same. But in Variant 2, there is nothing morally wrong in Sophie making either choice. She is unable to save both, and the choice is tragic, but nothing wrong is done, just as the person who witnesses one of her children being murdered and is unable to do anything about it because she is tied up and gagged has done nothing wrong. In both cases, there will be feelings of guilt, of course, but I suspect those will be mistaken to some extent.

In Variant 1, however, it is wrong for Sophie to make either choice. The right thing to do is to say nothing, or to say something unhelpful like "I want both to live." For it is wrong for her to lie and say she wants one of the children to die when she doesn't, and it is wrong for her to make herself want one one of the children to die. So the right decision there is the one that results in the deaths of both.

On the other hand, the captain who comes to the rescue after a shipwreck and decides whether to send his rescue boat to the starboard of the sinking ship where there are 300 people in the water or to the port of the sinking sip where there are 30 people in the water does absolutely nothing wrong in sending the rescue boat to the side with 300. The duty not to kill the innocent (perhaps apart from a divine command, God being the owner of life) is absolute. The duty to rescue the innocent from death is prima facie.

I am worried that while building a non-consequentialist case, you are allowing yourself to be swayed by consequentialist intuitions that conflate the foreseen with the intended, the permitted with the effected.

I could be wrong, but I thought St Paul was saying that one may not sin so that *grace* may abound. Is there some other passage about one not doing evil so that good in general might come of it? Why think the text in question provides sufficient support for the more general principle?

If it does, this strikes me as interesting, since it pretty definitively rules out at least some versions of consequentialism. Most versions of act-consequentialism will say that, in Alex's original example, my killing Jones is permitted, since it results in so many lives being saved. But then killing Jones isn't *evil* at all. In general, it's hard to see how any act that results in more good could be evil on straight-up act-consequentialist theories. So then it's hard to see how to make a sense of a general principle like "Don't do evil that greater good may come of it"--at least where the "may come of it" is cashed out in terms of the consequences of the act in question.

But it seems strange to me to think that St Paul intended to rule out, in one fell swoop, act-consequentialism (or something like it) w/ that one sparse stroke of the pen. I don't have an argument or anything; just a feeling that such an extrapolation requires a pretty fair bit of interpretive license, and that we might be in danger of overstepping. If Alex's interpretation of the text in question amounts to an out-and-out denial of act-consequentialism, I guess I'm more inclined to think that something's gone wrong in the hermeneutic-exegetical lifting than follow the interpretation where it leads.

Also, is felix culpa really that prominent in the tradition? My understanding was that it's a less-than-majority view, though I'm a bit out of my reach here.

Luke:

Paul says that some say that we Christians say: "Let us do/effect bad stuff that good stuff may come (PoiEsOmen ta kaka hina elthe(i) ta agatha)?" (Romans 3:8). (I am trying to keep all that is ambiguous in the Greek ambiguous in my translation.) It's clear that Paul thinks the sentiment that Christians are accused of having is wrong.

Yes, I think Paul, like the Christian and Jewish traditions, takes it as obvious that act consequentialism is false. That is why, for instance, one should be willing to have oneself die, along with one's family if that is how it has to be, rather than deny Christ, a judgment the tradition universally held.

I agree that if act consequentialism is right, then killing an innocent person is sometimes not an evil action. But such a killing is, nonetheless, the direct effecting of an unequivocally and intrinsically bad/evil thing--the death of an innocent person.

Hi Alex,

I'm sure you're right that the tradition has in general rejected act-consequentialism, though I wonder whether it's really seemed obvious to all members of the tradition that it's inconsistent w/ it. (No doubt counterexamples could be found!)

It seems to me that Rom. 3.8 is a tricky passage: it's a rhetorical question, that comes in the passage of a back-and-forth where it's not immediately obvious who the interlocutors are supposed to be. In fact, Douglas Campbell has an intepretation of Romans where the first few chapters are actually Paul's statement of his opponents case against him. This just to say that there's gonna be some textual ambiguity, and so--for anyone willing to question the tradition on these matters--ambiguity in general on just what 3.8 should mean for us today.

It seems to me that a Xian act-consequentialist could make sense of the injunction not to deny Christ, even if one's family is at stake, on the supposition that the individual stands to lose his or her eternal status, her place w/ God in the afterlife, by denying JC--which is arguably an evil that would outweigh the earthly death of one's family.

Luke:

Of course, once one brings in the afterlife, act consequentialism becomes a very hard to apply doctrine. For instance, the fact that killing one innocent person may save a billion innocent lives is no longer the decisive question. The decisive question, rather, is about the infinite good of heavenly life.

Now one might posit that (say, by divine decree) the infinite good of heavenly life is always more probable when one does the action that best optimizes all finite goods.

If we say this, then in the martyrdom case, heavenly life is more likely to come with the choice that optimizes the finite goods. Suppose we say this and then try to support the early Christian conviction that it is better to die with one's family than to lie about being Christian. We need to say that the disvalue of lying about being Christian is much worse than the deaths of several innocent people. If so, then this, together with act consequentialism, implies the following abhorrent judgment: If by killing several innocent people I could dissuade Jones from lying about being Christian, then I should kill these innocent people.

So I don't think we can bracket the heavenly goods. But if we don't bracket them, then the consequentialism is of no practical value.

All that said, that may be fine in the context at hand. One might, for instance, take the following variant of a suggestion of Jeremy's: Act consequentialism is the correct moral theory. God directly knows consequences, so he is an act consequentialist simpliciter. But we need to obey God rather than weigh consequences ourselves, because the most significant consequence, heavenly life, is one about which we know primarily by revelation.

I don't like this view, but it's not that terrible: it may be extensionally correct. :-)

Interestingly, though, Augustine at least expressly dismisses a consequentialism that does consider heavenly reward. He thinks that the unbaptized are not saved. But if a lie were requried to get into a jail to baptize someone, we should not utter that lie.

BTW, in her new book (and probably before; I can't remember), Frances Kamm rejects the claim that, if M is a means to E and you intend E, you intend M. She does this by playing around w/ Loop variants on the Trolley case, where I can prevent the trolley from hitting five workers only by directing it on to an adjacent five where one worker is placed, and the track loops around such that, were the one not there to stop the trolley, the trolley would loop around and kill the five from the opposite direction.

Suppose in this case hitting the one is a means to saving the five. According to Kamm, one of the tests for whether you, the driver, intend to hit and kill the one by turning the trolley is whether, were the one not there, you would turn the trolley anyways. If not, you intend to hit the one by turning the trolley. Kamm, however, thinks you could rationally turn the trolley even if the one wasn't there, since saving the five from the (non-looped) oncoming trolley is at least one of the things you need to do to save them outright. So you have at least *some* reason to turn the trolley even if the one isn't on the sidetrack.

Another test is whether, if we turn the trolley toward the one, we are prepared to take extra steps to ensure that the one is hit should this be needed; if so, we intend to hit him. But we need not be prepared to take extra steps, since we may see that, if we were to take further action to hit the one, this would involve intending to hit the one--which we may not be willing to do. Kamm even thinks that once we have turned the trolley we can rationally take steps to prevent the one from being hit; if so, this shows that we don't intend to hit the one.

This is all going to turn on the proper tests for intention. But it does show that there are well-articulated views on which it's false that one must intend the means to an intended end--which is, I think, just what Ted was saying above.

This said, I'm not sure how this bears on Alex's argument. Is there some story to be told about how God might pass the counterfactual test in allowing Adam to sin? I'm not sure I see how God could have sufficient reason to permit it if he knew, or was highly justified in thinking, it wouldn't result in greater good.

I should add that in the Kamm examples you intend to save the five.

Luke:

Counterfactuals can at most be a guide here, they can't really be determinative of intention. Counterfactuals are way too context dependent, and their semantics too unclear. :-)

"Kamm, however, thinks you could rationally turn the trolley even if the one wasn't there, since saving the five from the (non-looped) oncoming trolley is at least one of the things you need to do to save them outright. So you have at least *some* reason to turn the trolley even if the one isn't on the sidetrack."

It either is or is not the case that there is reason to turn the trolley if the one isn't on the sidetrack.

If there is no such reason, the example fails.

Suppose then there is a reason R to turn the trolley absent the one. This implies that if the one were there, it would be possible on account of R to turn the trolley towards him without intending to hit him. I agree: but then his death is not a means towards the end that one is pursuing. For the end that one is pursuing is the one embodied in R. Let's suppose that that reason is that five not die by a trolley coming at them from, say, the right (suppose that the unlooped trolley will come at them from the right, and the looped trolley from the left). Well, the goal of them not dying by a trolley coming at them from the right is not in fact served by the trolley hitting the one guy on the side-track. So his death is not a means to the end that you are actually pursuing.

But what if you think as follows: My goal is to save the lives of the five. To save their lives, they must be saved from being hit from the right and from being hit from the left. By turning the trolley even absent the guy on the side-track, I am saving them from being hit from the right, and hence I am promoting the end of saving their lives.

But as it stands, that would be silly. One does not try to prevent a disjunction p-or-q by ensuring that q rather than p happens.

OK, but maybe we can improve on this. Maybe the thought going through your mind is this: "If I shift the trolley, they certainly won't be killed from the right. They might or might not be killed from the left. So by shifting the trolley I am decreasing the probability of their being killed, and thus promoting the goal of saving their lives."

This is only going to work, however, if in one's view of the situation, the likelihood that they would die with the trolley continuing is greater than the likelihood that they would die with the trolley turning. But maybe that's so. Maybe you think: "The other way around is longer, so more things might happen." Or maybe you don't see it as clearly.

Now we are getting somewhere interesting. For it seems like we now have a reason R for turning the trolley absent the guy on the side-track, and yet where the end in view is the saving of the lives of the people on the track.

It seems that in this case, even if the guy were there on the side-track one could act on the same reason R, for the same end, which is the saving of the lives. But then it seems the death of the guy on the side-track would be a means to the end that one chose, and Kamm wins.

But not so. What we get then is that the death of the guy on the side-track promotes your end and is caused by you. This does not mean that the death of the guy on the side-track is a means to the end. For not everything that promotes your end and is caused by you is your means. First of all, something that you know nothing about can promote your end and be caused by you, and something that you know nothing about surely isn't intended. Secondly, just finding out an additional fact does not change your means. Hence, it is possible to know that something will be caused by you and will promote your goal, without it being your means.

(The standard example of a situation like this is the Dept Chair case. You are the Dept Chair and are afraid that Dr. Jones, a star, will leave the Dept. So you bestow an honor on Dr. Jones, to make him more likely to stay. You know that Dr. Smith, whom Dr. Jones hates, is likely to feel offended by the honor shown to Dr. Jones, and will leave. This further increases the chance Dr. Jones will stay. But it need not be the case that Dr. Jones' staying is your means.)

So what Kamm has perhaps provided is a counterexample to the principle that if X is caused by you and is known by you to promote your end, then X is an intended means.

(I worry a bit that on my view "means" means "chosen means". But I am fine with that--it's mainly a question of what your plan of action is.)

"So what Kamm has perhaps provided is a counterexample to the principle that if X is caused by you and is known by you to promote your end, then X is an intended means."

Suppose we replace 'caused by you' w/ 'allowed by you' (which seems to me unobjectionable). Then doesn't the principle show that it no longer follows from the fact that God allows Adam's sin so that grace may be manifested (or whatever) that God intends Adam's sin as a means to this end? (Assuming that God knows the upshot of Adam's sin--but I take it we're assuming as much.)

Dear Luke:

No, because of the "so that". The mere fact that Adam's sin promotes great goods intended by God does not, indeed, entail that God intended that sin. But when one adds that God allows the sin because of the great goods, then the conclusion is much harder to avoid (though my solution comes close). In the trolley context, if someone turns the trolley because of the guy on the side-track, then surely he intends to hit that guy.

Hi Alex,

But I think that's just what Kamm's ultimately arguing for: a distinction between because of/intends. She thinks we can do y because it results in x, without intending to do y.

Another Kamm example: Suppose you want to throw a party but are strongly averse to cleaning up the mess you know will result. You also happen to know, however, that your friends are the type of people who will help you clean up. If you had to clean up yourself, you wouldn't throw the party. But since you know your friends will help, you decide to throw the party.

Here Kamm thinks you throw the party (at least in part) *because* you know your friends will help with the mess. But it seems strange to think that your intent in throwing the party is that your friends clean up the mess that results. You throw the party intending to have a good time.

Dear Luke:

This case roughly has the structure of my proposed solution to the original conundrum. There is an end, say the conversation at the party (let's suppose these are philosophers), which provides one's reason for the action. There is a defeater to the action: the hard work of cleaning up. There is a defeater to the defeater: the help you will get from the other people.

I think these kinds of examples only have a hope of working when the non-intended consequence is only a part of the reasons for the action, so that the action would not make sense as done only for that consequence. It would make no sense to have a party just to have a joint cleanup (unless, say, one believes that joint work is good for the body and soul; but then the joint cleanup is clearly intended).

A primary reason justifying the action is still needed. Recall my proposed solution to the felix culpa problem: There is a primary reason for allowing Adam to sin, namely allowing him to have free will. The redemption that God would work after Adam's sinning is not the primary reason for God's allowing the sin, but serves as a defeater to a defeater.

All that said, I am worried about Kamm's example, in a way in which I am not worried about my case. Take the action-failure test (this is a rough formulation of something I've learned from a paper by Mark Murphy): if an action is expected to cause E, and if one's plan of action would at least in part be a failure were E not to occur, then E is intended.

Now, God's plan would in no way be a failure if Adam and Eve and their descendants were not to sin. True, the redemption would not occur. But the redemption was not a part of God's plan simpliciter, but contingently on sin.

However, my plan for the party would indeed in part be a failure if the guests did not chip in with the cleanup. So their chipping in is a part of the plan, and hence intended. (In the end, I care more about the notion of an action plan than that of an intention.)

I don't have time right now to read everything that follows Dominic's comment, but his immediate impression is wrong. The view I've been describing is not rule consequentialism.

Rule consequentialism does have a level distinction between the consequentialism at the foundational level and the rule-following of the level of what makes actions right and wrong. But rule consequentialism as a moral theory would apply to every moral being, including God. On the view I was outlining, act consequentialism would be the best moral theory for God to use.

I think there are other differences too, but I'd have to think that through.

Just to possibly widen the discussion, the felix culpa is pretty much non-existent in the Eastern Orthodox tradition.

You might find some of the writings of David Hart on Evil here as closer to an accurate representation of the Eastern tradition than Hick's.