It’s among the importantly neglected problems in philosophical theology that we have no success conditions for theodicy. What conditions must a theodicy meet in order to succeed? I want to approach this problem by considering some (possible) challenges to successful theodicy. I take theodicies to be consistency proofs of a sort. They show that the existence of God (a morally perfect being, a perfectly loving being, etc.) to be consistent with the kinds of evils we actually find in the world. Perhaps ‘thicker’ theodicies will invoke theological doctrine, but I’d rather avoid additional assumptions if possible. Would a theodicy that made any of the following assumptions fail? Must theodicies make one or more of the following assumptions?
(i) There are goods that outweigh the terrible suffering of children, but it’s never morally permissible to allow the suffering of a child as a means to a greater good. And no theodicy can avoid assuming God is doing this.
(ii) There are goods that outweigh the terrible suffering of children, but it’s never morally permissible to allow the suffering of a child as a means to a greater good for others. And no theodicy can avoid assuming God is doing this.
(iii) There are goods that outweigh the terrible suffering of children, but it’s never morally permissible to allow the suffering of a child as unavoidable to a greater good. And no theodicy can avoid assuming God is doing this.
(iv) There are no goods that outweigh the terrible suffering of children, and it is never morally permissible to allow the suffering of a child as a means to a lesser good. And no theodicy can avoid assuming God is doing this.
(v) There are goods that outweigh the terrible suffering of children, but it is never morally permissible to intend the suffering of a child as a means to a greater good for itself or others. And no theodicy can avoid assuming God has such an intention.
(vi) There are no goods that outweigh the terrible suffering of children, and it is never morally permissible to intend the suffering of a child as a means to a lesser good. And no theodicy can avoid assuming that God has such an intention.
(vii) Not even the state of affairs in which 500 children are prevented from suffering E outweighs the disvalue of one child’s terrible suffering E. And every theodicy must assume that there are goods/worse evil that outweigh each actual evil.
(viii) There are no goods G such that for every actual evil E, G entails E and (G + E) > (~G + ~E). And every theodicy assumes that there are outweighing goods that entail actual evils.
(ix) For every actual evil E and good G such that G entails E and (G + E) > (~G + ~E), there is a greater good G’ incompatible with G such that G’ does not entail E. And every theodicy assumes that for every good that entails an actual evils there are no greater goods that do not.
Mike: Thanks for this post. That's quite the list you've drawn up. Your characterization of theodicies as "consistency proofs of a sort" seems to fit what Plantinga calls "defenses" better than it does theodicies. At least as I see it, theodicies must explain why God permits this or that evil, or the world's evils in general, and consistency with the data is only the first (and weakest) ingredient in a good explanation.
Some of the items in your list mention trade-offs that require God to allow (or perhaps even intend) a child's suffering. Any good theodicy -- any good explanation -- invoking such trade-offs must, I think, explain how they constrain an omniscient and omnipotent God, both in general and with regard to the particular case(s) of evil being explained.
Hey Steve,
I guess I was thinking of this sort of consistency proof as a theodicy since it must show that God can coexist with actual evil and any successful attempt to show that would also show, as far as it can be shown, that the suggested reasons for actual evil are the actual reasons. This could be made clearer. I have in mind finding a world that duplicates our world but includes God and justifying reasons for all of the evil.
Explain to me what you mean by 'explaining evils'. Suppose God would have to 'violate' a causal law to prevent an evil, does that constrain him? No one can violate a causal law, not even God, since I take it that minimally laws are exceptionless regularities. Or are you saying that any explanation would require the logical/metaphysical impossibility of God preventing E (or of God getting G and preventing E where |G| > |E|)? Would it explain the fact that I allow you to cause pain to Smith that I accepted the deal that you pay me $10.00 and I allow that?
I do think there is a serious worry for God allowing or intending evil for the sake of a greater good, and I do think many theodicies are morally sloppy in this way.
Explain to me what you mean by 'explaining evils'.
Roughly: Identifying a morally sufficient reason for God's having permitted one or more particular evils, where 'God' denotes a perfect being.
Suppose God would have to 'violate' a causal law to prevent an evil, does that constrain him? No one can violate a causal law, not even God, since I take it that minimally laws are exceptionless regularities.
I don't know if it's significant that you enclosed 'violate' in quotation marks the first time, but I'll presume not. I'd ask the theodicist to explain why God couldn't prevent the evil without violating a causal law. Furthermore, this theodicy seems to be saying that not even God can violate a causal law because causal laws are defined as exceptionless regularities. But that's just a tautology: not even God can produce exceptions to exceptionless regularities. If that's the way 'causal law' is being defined, then appeal to causal laws doesn't explain why God didn't prevent the evil.
Would it explain the fact that I allow you to cause pain to Smith that I accepted the deal that you pay me $10.00 and I allow that?
I'm not sure how this question is relevant. It certainly wouldn't explain God's allowing Smith to suffer at my hands, because it wouldn't explain why a perfect being allowed it. (In keeping with your post, let's assume that Smith is an innocent child: Suzy Smith, age 5.)
Thanks Steve,
The last question was aimed at trying to elicit your ideas about a particular moral view; it is not yet directly related to what God can or can't do or allow. So, it goes something like this:
a. It is permissible for me to allow my child to undergo severe suffering E which is caused by several surgeries attempting to improve her condition even if the results are not not very good.
b. It is not permissible for me to allow my child to undergo torture E (same as above in quality and quantity) in Auschwitz even if I am correctly promised that the torturers will afford my child and others extremely great goods afterward (goods whose absolute value outweigh the disvalue of the suffering)
Lots of people seem to think that (a) is true and (b) is too. I wonder about why, and for a moment I thought it might have to do with the suffering in (a) having an explanation of evil (of the sort you're looking for) that the suffering in (b) lacked. But I think I gather that the suffering in (a) and (b), by your lights, have sufficient explanation. Is that right? So, even if the suffering in (b) is necessary to the greater good, and has the sort of explanation you're requiring, many think (and I'm pretty sure I think) it is not permissible. So I'm at a loss as to why I think (b) is true, though I'm pretty sure it is.
Furthermore, this theodicy seems to be saying that not even God can violate a causal law because causal laws are defined as exceptionless regularities. But that's just a tautology: not even God can produce exceptions to exceptionless regularities. If that's the way 'causal law' is being defined, then appeal to causal laws doesn't explain why God didn't prevent the evil.
I'm sure I'm missing something. Let world W include causal law L, good G and evil E, where |G| > |E|. Suppose it is true in W that good G causally entails E, so there is no way to prevent E without violation of physical law L or loss of G. Or, producing G in W causally requires E. When asked, why God didn't prevent E in W, the answer is that God cannot violate L and would have had to violate L to prevent E. Why isn't that an explanation? If you ask Smith why he didn't fly at slightly faster than the speed of light to save Jones, why doesn't it explain his not doing so that it would have violated a causal law?
Incidentally, that causal laws are exceptionless regularities is not a definition I'm introducing. It's just what causal laws are at a minimum, on any further account of what causal laws are; they might also be metaphysical necessities. But I'm not taking any stand on that here.
a. It is permissible for me to allow my child to undergo severe suffering E which is caused by several surgeries attempting to improve her condition even if the results are not not very good.
b. It is not permissible for me to allow my child to undergo torture E (same as above in quality and quantity) in Auschwitz even if I am correctly promised that the torturers will afford my child and others extremely great goods afterward (goods whose absolute value outweigh the disvalue of the suffering)
Concerning (a) and (b), I'm trying to address one of the worries in the literature on the possibility of theodicy. The worry goes this way: It is not permissible for God to allow the severe suffering of children even in cases where doing so is necessary to a much greater good. And this is because doing so is analogous to (b). It is analogous to a parent allowing a child to be tortured for a greater good. The worry here, I think, is that, unlike the case in (a), the evil in (b) is extorted for the greater good. It is not permissible to allow evils to be extorted even for a much greater good. The evils in (a) are not extorted.
Mike:
These are good and hard questions.
In regard to (i)-(iii) and (v), I think it's good here to reflect on stuff from the Principle Double Effect (PDE) literature, especially Kamm's triple effect cases which appear to show that a good can play a justifying role for an action even though that good is not intended either as a means or as an end. In other words, Kamm cases open the possibility for there to be cases of a good G playing a role in justifying allowing an evil E, without E being allowed as a means to G.
Imagine a compound that contains an evil scientific institute that tortures children in order to learn more about pain responses. Indeed, it has hundreds of children in the compound. You are a police officer hidden in the bushes outside the compound, making early preparations for an operation that you know would save these children. The preparations for the raid will not be complete for another eight hours.
Suddenly, you notice that one of the evil scientists is torturing a child, alone in an outlying shed. You know from experience that the child will be dead in eight hours. Listening in on cellphone conversations, you also learn that some time after the scientist is done with torturing this child, they will evacuate all the children to a more secret location in another country by airplane (it's on a landing strip on the institute grounds), and then they will be torturing the children at their leisure, with no hope of rescue. You could rescue the child right now. If you do that, however, the members of the institute will pack up and leave, and many children will suffer. While if you don't rescue the child, the raid will go on as planned in eight hours, just in time to rescue the remaining children. You have excellent reason to think the raid will be successful.
Note that there are two variants of the case, and one might have different judgments about them:
Variant A: Rescuing the child causes the evil scientists to leave with the remaining children by warning them about an impending raid.
Variant B: Rescuing the child causes the evil scientists to leave with the remaining children because the only thing they were waiting for before leaving was completing the torture of the child.
I'll get back to the two variants. In any case, here comes Double Effect. You reflect on the fact that rescuing the child will be a non-intentional cause of suffering among many children, by causing the children to be taken to a different location. By Double Effect, it is only permissible to act in a way that causes an evil if the evil is proportionate to the good. The good you would be striving for is to save a child from suffering. But the evil of so very many other children suffering is disproportionate here. So Double Effect forbids you from preventing the torture of the child.
So you don't rescue the child from torture. But now it seems that you allowed a child to be tormented for the sake of saving many other children from suffering.
What can someone friendly to the deontological principles associated with PDE say about allowing the torture to continue, by refraining from rescue? I see three main options:
1. The good of saving many children can justify your allowing the torment of the one child when--as in this case--you do not intend thereby to save these children and you are not allowing the torture as a means to saving them. This is a Kamm-style answer (though the case is mine, and I don't know for sure if she would go with this line in this case).
2. You are allowing the torment of the one child as a means to saving many, but it is permissible to allow, but not commit, an evil as a means to a proportionate good (many children being saved from suffering). In other words, deontological principles like PDE apply to commission but not to omission. This fits with some people's intuitions about differences between passive and active euthanasia.
3. This case is a real moral dilemma. On the one hand, PDE forbids you from rescuing the one child, since rescuing the child will cause a disproportionate evil. On the other hand, PDE applies to allowings, and hence forbids you from refraining from firing the warning shot, since the refraining would be an allowing of an evil for the sake of a good. So you do wrong whether you rescue or don't rescue.
Given options 1 and 2, I don't think we should go for 3. Recourse to real moral dilemmas should be one of the last resorts in moral theorizing. Moreover, especially in Variant A, it seems very plausible to me that refraining from rescuing is the right action.
Now if we go for 1, then that opens the theodical option that God is justified in allowing the suffering of children in the light of a great good, but without his intending that suffering (or even the allowance of that suffering) as a means to that good.
I see these kinds of cases as defeater-defeater cases. You have a defeater to your plan to refrain from rescuing: the defeater is that a child will suffer. But you have a defeater to this defeater: many more more children will suffer if you do rescue.
And if we go for 2, then that opens the theodical option that God is justified in allowing the suffering of children as a means to a great good. He would not have been justified in producing that suffering as a means to a good, but allowing is morally different from producing.
Note that one might morally distinguish between allowing-the-torture-to-continue being a means to saving children and the-torture-continuing being a means to saving children. If so, then one might say that in Variant A, it is the allowing that is a means to saving the other children, but the torture itself is not, while in Variant B, it is the allowed torture that is a means to saving the other children. I don't know how this affects matters morally. In any case, the difference could be theodically relevant. For there are cases when it is the evil itself that makes possible a good, and there are cases when it is the allowing of the evil that makes possible a good. (It is not the evil of choosing wrongly that makes possible the good of choosing in a significantly free way, but it is the allowing of the evil of choosing wrongly that makes it possible. On the other hand, it is the evil of suffering unjustly that makes possible certain cases of forgiveness.)
Thanks very much, Alex. This is interesting. Is the Kamm stuff in her _Intricate Ethics_? I think I might have misunderstood one thing about the case you describe. Doesn't not rescuing the child and waiting for the raid result in the tortured child dying? The raid won't occur for 8 hours and in that time the child will be dead. If that's so, then I think you cannot wait. But I want to say something about the way you develop the case. It is much easier (though of course extremely difficult) for us to make a moral decision here than it would be for God. Afterall, God set the whole thing up, knowing that he had the option never to create any of the torturers or the circumstances. So, my worries involve the permissiblity of allowing or intending evil in the way God might have to in such cases--creating these sorts of torturers and placing them in these sorts of circumstances--when there is the option of having created others or of placing the would-be torturers in other circumstances. Even if there is some big payoff God sees for creating these torturers and putting them in circumstances that he knows will lead to the torture of innocent chidren, many might say (I might say) forego the big payoff.
So I guess I want to get you to think about how God might create a world like ours in which there are the sorts of torturers and suffering you describe. I don't ask that as an open ended question. I ask how that very act, the act of creating such beings and putting them in circumstances in which he knows they will cause such suffering to children, is consistent with moral views we find highly credible. Is there a way of understanding that act that makes it plausibly moral? You mgiht compare a trolly case in which you get to decide whether anyone will be on the tracks at all, in addition to decidng whether to switch the lever from five to one, if you do choose to have people one the tracks.
Mike,
It was by mistake that I posted my previous comment on Helen’s thread. I am sorry about that.
You write: “A successful theodicy would have to explain the existence of actual evil, why we have the kinds and distribution of evil we have.”
Quite right. But also a successful theodicy should not start with a false assumption, or a false mindset. We must approach the problem from God’s point of view.
Suppose that God has one, undivided, and overwhelmingly good purpose for creation. Suppose further that that purpose entails the creation of many imperfect persons in a condition which includes a great number of individual goods but also some individual evils. Suppose, finally, that God’s respect towards these created persons (or, if you will, God’s justice) entails that God would create the world in such a way that individual goods and evils would be distributed *randomly* among them. Let’s call this the “undivided purpose divided creation” assumption, or UPDC for short. (Please observe that there is no tension between UPDC and the actual kinds and distribution of goods and evils. On the contrary, UPDC says that goods and evils are randomly distributed, as in fact they prima-facie seem to be. But to actually explain the kinds and distribution of goods and evils one first has to find out what God’s purpose for creation is.)
Now I’d be happy to discuss why I think that UPDC is true. But for now let us only assume that it is true. If it is true then it is misleading to ask how some particular evil is justified. We humans living in a divided universe are used to counting bits and pieces, and thus imagine that God does the same, spreading out individual goods and evils on a ledger as it were, and making certain that each individual evil is justified by outweighing individual goods which wouldn’t be possible unless that evil obtained. It is well known that going down this path greatly complicates matters, for there are a great number of cases that must be answered individually. Whereas, on UPDC, any particular evil obtains for no particular reason at all – it simply belongs to a good world in which such evils may randomly obtain. (Incidentally, to answer the question of why some particular evil obtains is as difficult as to answer as the question of why some particular good obtains. The atheologian might as well construct an argument from the unjustified distribution of good.)
We don’t know that UPDC is true. My argument though is that we should think about UPDC and try to find something wrong with it before going down the complicated path of discussing theodicy on the assumption that UPDC is false. And I observe that theodicists don’t typically do that, but rather jump right in and discuss the many problems of the ledger-type view. I suppose philosophical tradition plays a key role here. Traditionally theists have interpreted God’s perfection quantitatively, and indeed as entailing infinite quantities (e.g. God knows all that can possibly be known), and thus assumed that God must be micromanaging every last minute bit of creation. Perhaps a lighter and brighter understanding of God is closer to the greatest being we can conceive.
Now I’d be happy to discuss why I think that UPDC is true. But for now let us only assume that it is true. If it is true then it is misleading to ask how some particular evil is justified.
I'm not sure why you say this. Suppose that S suffers E as a result of a random distribution of evils. We immediately have the worry that, had God prevented S from suffering E, there would have been an overall improvement in the world. In order to argue that there would not be an overall improvement were S prevented from suffering E, you'd have to show that there is a precise amount of evil necessary to a greater good and such that it does not matter who suffers the evil. But there is zero reason to believe this.
Apart from that, you're approach to the problem is way out ahead of the concern. The concern is that theodicy is not possible. And I offered some reasons why no theodicy could succeed. So my worry is about the modal claim that there can't be a successful theodicy. And the reason there can't be such a theodicy is because all theodicies entail that a perfectly moral and loving being might intend the suffering of innocent children for the sake of a greater good. But no moral being can do that. This is what Alex was addressing in his comment. So, what you'll need to show me is how God might choose to create a world in which children are caused to suffer in the way described, presumably for the sake of some greater good, when God might have actualized a less good world with less suffering in it, or might have actualized a world with no sentient beings at all in order to avoid that specific kind of suffering, or might have chosen a thousand other options that did not involve that sort of suffering.
How come guys neglect to take into account the Fall of Satan?
Plantinga does take the (possible) work of evil spirits into account in his free will defense. It does not help with the problem of the possibility of theodicy.
Evil is just an effect of the Fall of Satan, and the autonomy Satan gained by the performative act of DECLARING WAR to the Jewish People and, by proxy, the whole humanity.
Or should a theodicy explain the Fall of Satan itself?
Mike,
“Suppose that S suffers E as a result of a random distribution of evils. We immediately have the worry that, had God prevented S from suffering E, there would have been an overall improvement in the world.”
Had God prevented S from suffering E, the distribution of evils in the world wouldn’t be random. But on UPDC’s view God’s purpose for creation entails that they be random. Thus, God preventing S from suffering E would defeat God’s purpose for creation. (This is putting matters roughly. The question is not really about God preventing one S from suffering one E, but rather that on the principle that God would in fact prevent that one S from suffering that one E, it seems that God would also in fact prevent many S’s from suffering many E’s. Which, on the assumption that God’s purpose requires that the distribution of evils (and goods) be essentially random, does defeat that purpose.) But, since God’s purpose for creation is good, God will clearly not defeat that purpose. To do so would be both immoral and incoherent.
“all theodicies entail that a perfectly moral and loving being might intend the suffering of innocent children for the sake of a greater good”.
The very idea of God *intending* somebody to suffer strikes me as incoherent. Never mind the suffering of innocent children, and never mind for what greater good’s sake. Evil and good are not coins to be charged against each other. All evil is absolutely alien and horrible in the eyes of the greatest being I can conceive, and is never used for whatever reason. Now please observe that on UPDC’s view that the few evils (and the many goods) in creation are distributed randomly, it is not the case that God intends any one of them. Rather, given that temporary and random evil is entailed in God’s good purpose for creation, God suffers it, and on the Christian view suffers it completely.
Had God prevented S from suffering E, the distribution of evils in the world wouldn’t be random. But on UPDC’s view God’s purpose for creation entails that they be random. Thus, God preventing S from suffering E would defeat God’s purpose for creation
This is just question-begging. The question is whether we have any reason to believe UPDC si an acceptable principle in the first place, and we don't.
I don't see the discussion as making much progress. As I mentioned, the fundamental problem concerning the possibility of theodicy for the reasons offered above, in a few posts. That's more or less the problem I'm interested in.
I think there is reason to believe UPDC is an acceptable principle.
On the view that theodicy is a consistency proof, the very fact that UPDC removes many of the problems theodicists struggle with is one such reason.
But I think a full-blooded theodicy wants to show that given God’s perfection one can see that God would want to create a world like ours. Here theodicy becomes an entailment proof. Now, one such reason for UPDC is to see that God’s love and respect for each human being entails that God would not want to distribute individually the many goods and some evils of creation. As an analogy consider a loving parent who wants to give an unearned present to one of her children. Rather than choosing herself which the favored child will be, she would rather let chance decide.
Our discussion here moved me to see another reason. Ivan Karamazov would refuse even the kingdom of God if it had to come at the price of one child’s suffering. I think this is the right relationship between a moral person and a particular evil, namely absolute and unconditional rejection. With more reason I hold that God would never stoop down to using an evil for a good purpose. The very idea is close to incoherent, for using an evil for a good purpose makes it good. In the same sense God would never choose an evil, or intend an evil to obtain. On the other hand, as I think everybody agrees, there are great goods for humans which cannot obtain at the absence of evil. The obvious (and it seems to me only) solution is for God to create a world which produces goods and evils in a random fashion.
I think there is reason to believe UPDC is an acceptable principle. On the view that theodicy is a consistency proof, the very fact that UPDC removes many of the problems theodicists struggle with is one such reason.
As I said a post or so ago, UPDC has an obvious problem. If S suffers E as a result of the distribution D of evils recommended by UPDC, then there must be no other distribution D' such that there is less evil suffered under D' and the world is overall better under D'. But obviously there is such a distribution. It is the one under which S does not suffer E. S's suffering is unnecessary to any greater good. So, UPDC is not up to the task of resolving any theodical problems. Your response so far has been "preventing S from suffering E would yield a non-random distribution". That's exactly the point. UPDC yields the wrong distribution of evils if one is concerned with theodicy. It yields distributions in which people suffer needlessly.
Perhaps I gave the impression that the UPDC principle is a full-fledged theodicy. I only think that the UPDC is a principle which a successful theodicy will include. But the UPDC itself is not a theodicy. In particular it does not justify the amount and types (i.e. distribution) of goods and evils in creation, i.e. why God would want to create a world with that distribution. It only observes that *if* that distribution is justified then it does not follow that every single good or evil must also be justified. Perhaps it is the case that there is justification and reason for God to create a world with our distribution of goods and evils, but in which single goods and evils have no reason or justification, and obtain randomly. Hence, the UPDC says that this is how things stand: the world as a whole is justified but not every single good or evil in it. And the fact that the UPDC principle removes from theodicy one difficult problem (namely the need to justify each single good and evil) is one reason why one should consider embracing it. So that’s my whole claim: Theism does not entail that each single good or evil which obtains in our world must have a particular reason and justification – as most people who think about theodicy assume.
Now the problem of what justifies the distribution of goods and evils in the actual world remains. In this context I would like to suggest the reasonable limits of the theodicist’s burden of proof. I think the conditions that a successful theodicy must meet is to show 1) that a perfect being would have purpose P for creation, 2) that P entails the creation of a world in which not only goods but also some evils obtain, and 3) that what we know about the actual distribution of goods and evils in our world is consistent with P, or, conversely that after lots of thinking one cannot find a defeater against that consistency.
Finally, I’d like to comment on the following:
“If S suffers E as a result of the distribution D of evils recommended by [some theodicy which entails UPDC], then there must be no other distribution D' such that there is less evil suffered under D' and the world is overall better under D'. But obviously there is such a distribution. It is the one under which S does not suffer E. S's suffering is unnecessary to any greater good.”
On UPDC, S risks suffering E, and that risk is defined by the distribution D. Is a world with distribution D' where S’s risks are lower a better world? Perhaps it is, but then if one iterates the same argument one will reach at some point the optimal, where lowering even further the risks produces a world which is overall worse (as measured against God’s purpose for creation). The question then is whether given God’s purpose the distribution of goods and evils in the actual world is at this optimal point or not. And that’s the third condition of a successful theodicy I described above.
Perhaps it is the case that there is justification and reason for God to create a world with our distribution of goods and evils, but in which single goods and evils have no reason or justification, and obtain randomly.
Right, so what we need is some reason to believe that's true. I don't see one.
On UPDC, S risks suffering E, and that risk is defined by the distribution D. Is a world with distribution D' where S’s risks are lower a better world? Perhaps it is, but then if one iterates the same argument one will reach at some point the optimal, where lowering even further the risks produces a world which is overall worse (as measured against God’s purpose for creation).
Yes, and the atheodicist will say you reach optimality when there are no evils at all. None of them are necessary in the relevant sense to any greater good. The burden on the theodicist is to show that all of the actual evils are in fact necessary in the relevant sense. More to the point, and more exactly, to show that some of the most outrageous forms of suffering among the innocent is both necessary in the relevant sense and morally allowable. Not all of the evil that is necessary meets the standard of being allowable morally.
Mike:
"Doesn't not rescuing the child and waiting for the raid result in the tortured child dying? The raid won't occur for 8 hours and in that time the child will be dead."
Yes.
"If that's so, then I think you cannot wait."
This is a pretty controversial consequence, though.
And is the case I gave, at least in Variant B, significantly different from this one? You see a child drowning now in a pond whom you can rescue with no danger to yourself. But you know that there are ten children who will be drowning in ten minutes (maybe they are in a boat and you know that a malefactor has rigged the boat to fall to pieces in ten minutes) in a more distant pond, and if you rescue the one child, you won't have time to go to the other pond to rescue the other children. But if you refrain from rescuing the one, you will have time to rescue the ten. You have no special roles vis-a-vis any of the children.
In this case, it seems clear that you should go rescue the ten. After all, mere temporal ordering should not have all that much moral significance, and if the two drownings were simultaneous, you should obviously rescue the ten.
"It is much easier (though of course extremely difficult) for us to make a moral decision here than it would be for God. Afterall, God set the whole thing up, knowing that he had the option never to create any of the torturers or the circumstances. So, my worries involve the permissiblity of allowing or intending evil in the way God might have to in such cases--creating these sorts of torturers and placing them in these sorts of circumstances--when there is the option of having created others or of placing the would-be torturers in other circumstances."
All I wanted to point out is that there are ways to make great goods count in favor of allowing an evil in such a way that there be no clear violation of deontic principles, either because the evil is unintended or because perhaps it is permissible to intentionally allow an evil, though not permissible to intentionally produce an evil.
Now there are two questions:
1. Bracketing the deontic point, are there plausible theodicies?
2. If so, can they be made to work without violating plausible deontic principles?
I think the answer to 1 is affirmative, but that isn't the question at issue in our discussion right now. As for 2, we can look one by one at various proposed theodicies and see what we can make of them deontically. We now have at least four potential tools to use there:
A. Principle of Double Effect
B. the distinction between being justified by a good and acting for the sake of the good
C. the distinction between allowing versus doing
D. the differences between divine and human duties.
I think that between these four tools, many, though not all, of the proposed theodicies can be formulated in a way that avoids deontic worries. (One that I think can't be formulated that way are Calvinist theodicies on which God intends Adam's sin for the sake of greater goods. The Calvinist will lay stress on D, but I think that won't cut it.)
All I wanted to point out is that there are ways to make great goods count in favor of allowing an evil in such a way that there be no clear violation of deontic principles, either because the evil is unintended or because perhaps it is permissible to intentionally allow an evil, though not permissible to intentionally produce an evil.
Yes, I took you to be doing this, and I think you describe cases in which a reasonable person could come to your conclusion. My point was that I think we can't avoid a violation of the deontic principles in God's case, given the moral situation God is in. Surely God must choose among (1), (2) and (3) below. And there is no theodicy that can plausibly claim that God has opted for (3). He must have opted for (2). The difficulty in getting God off the deontic hook is that he foresees that the agents he freely creates (i) he need not have created, (ii) he need not have placed in circumstances where they can violate basic principles of justice and (iii) will engage in violations of basic principles of justice. If God freely creates beings and places them in circumstances where he knows will violate basic principles of justice, and he could have done otherwise, then he is on the deontic hook for such violations. In effect, he is not preventing serious injustices he might have prevented simply by not creating the relevant people or not placing them in the relevant circumstances.
1. I should actualize the most valuable world I can and create value-producing agents that always produce the most value they can.
(i) So long as they're producing the greatest amount of value, I'm justified in allowing them to inflict lots of suffering and freely violate basic principles of justice.
2. I should actualize the most valuable world I can and create value-producing agents that would produce more on balance value than any others I could create.
(i) So long as they're producing on balance more good than evil, I'm justified in allowing them to inflict lots of suffering and freely violate basic principles of justice.
3. I should actualize the most valuable world I can, and the best, constrained value-producing agents I can.
(i) I should never create a moral agent that will freely violate basic principles of justice, even if they produce on balance a great amount of value.
Mike,
How about this: Let the "deontically intendable" value of a world be the on balance value of the world insofar as it does not depend on evil for its actualization. Thus, suppose Sam is choosing between doing an evil E and a good G. If he does E, then a good G2 eventuates, and if he does G, then a further good G1 eventuates, with G1 less valuable than G2. Suppose he does E. Then the deontically intendable value of the world won't include G2, but will include G1, since getting at least as much good as in G1 doesn't depend on the evil E.
In worlds with no evil, the deontically intendable value equals the value.
By the Principle of Double Effect, if the evils in a world are not disproportionate to the world's deontically intendable value, the world is at least prima facie permissible to create.
Note that the value of freely choosing between right and wrong actions is great and deontically intendable, since this value does not depend on evil, even if the wrong action is in fact chosen.
Likewise, the value of beautiful and elegant arrangements of laws and initial conditions are deontically intendable, even if they lead to significant suffering.
Moreover, deontically intendable value isn't all that matters. For by triple effect kinds of considerations, there are values that can permissibly play a role in justifying actions that are not intended in the action. So we have a more general category of "justifying value" with respect to an action. Not all value is justifying value. If to save ten I must kill one, the lives of the ten, are not a justifying value with respect to the killing.
This is just setting up the problem. Now the problem of theodicy is to find the actual justifying values...
Thus, suppose Sam is choosing between doing an evil E and a good G. If he does E, then a good G2 eventuates, and if he does G, then a further good G1 eventuates, with G1 less valuable than G2. Suppose he does E. Then the deontically intendable value of the world won't include G2, but will include G1, since getting at least as much good as in G1 doesn't depend on the evil E.
That's clever, but why think deontically intendable worlds are permissible to actualize. Here's a counterexample, let G1 = G2. Getting at least as much as G1 does not depend on performing any evil action, so why would it be permissible to actualize a world with E? I see the point of the move, which tries to make it clear that E is not necessary to a certain quantity of value. From this I think you want to conclude that, in actualizing a world with E, you do not intend E as a means to G. I think that's false, but it's nonetheless thought provoking. Here's a counterexample to the principle you describe.
By the Principle of Double Effect, if the evils in a world are not disproportionate to the world's deontically intendable value, the world is at least prima facie permissible to create.
Suppose it is true that, if I push the man onto the tracks in front of the trolly we'll get the great good of saving 5 lives. Let that be world w. Am I permitted to do that? I think it is clear that DDE says no. But what if the deontically intendable value of w equals those 5 lives?! Let's make that true. You inform me that I could have saved those 5 lives without pushing the man into the trolly, so the evil is not necessary to the good of 5 lives. Maybe you tell me that all I needed to do was to switch the lever. Does it follow that it was permissible to push the man onto the tracks in front of the trolly? I'm pretty sure DDE would say no, but your quoted principle says it is permissible. The principle has to be mistaken, I htink.