Here’s my claim: there exists no gratuitous evil. The central condition on the existence of gratuitous evil E is that God could prevent E without preventing a greater good G (or permitting a worse evil E’). Consider a world W in which it is true that, for every instance of evil E, there is some individual I such that I responds with R to E and the value of R & E is positive.
W contains no gratuitous evil, since for every E there is some R such that □(R ⊃ E). Each response R to evil E is such that R is essentially a response to E. If God prevents E, then he prevents R, and R & E is more valuable than ~R & ~E.
Now I’ll make three claims about the actual world.
For every instance of evil E, it is possible that, for some individual I such that (I ≠ God), I responds with R to E and the value of R & E is positive.
It is true that, for many instances of evil E, there is no I such that I responds with R to E and the value of R & E is positive.
There is no evil E such that E is evidence that God does not exist.
But why is (3) true? We know that, for every E, there is some possible I such that I responds with R to E and R & E is positive. So, we know that for every E there is some possible R such that □(R ⊃ E) and R & E is positive. But then God cannot prevent E without preventing a greater good R. But then E is not evidence that God does not exist, though it is true that no I responds to E with R and R & E is positive.
[Updated 2.25.10]
I think the modal scope of (1) should be changed, as follows:
(1b) For every instance of evil E, there is some individual I such that (I =/= God) and it is possible that I responds with R to E and the value of R & E is positive.
Currently, (1) allows an evil to be nongratuitous when there is some merely possible individual who could respond to the evil with a greater good. So on this reading, it would be nongratuitous, for example, for God to cause an evil for which the only possible R is an act performed by, say, a talking unicorn; even when God knows full well that the actual world never did or will contain such a being. That certainly sounds gratuitously evil to me.
Instead, (1b) captures the idea that some actual individual is such that possibly she responds with R.
Just a quick note, though I'm not sure if this affects the main point. I'll say "R compensates for E" if R is a response to E s.t. R&E is a net good. So I gloss you as making this inference (starting at "We know that"):
Possibly, for each evil E, some R compensates for E.
So: For each evil E, possibly some R compensates for E.
I'm not sure how you're justifying switching the scopes here. Couldn't it be that, although at some worlds each evil has a compensating response, that's just because those worlds have less severe evils than ours? Claim 1 is compatible with there being an actual evil E which is "uncompensable"—i.e., it is essential to E that no response compensates for E. And it doesn't seem implausible that there really are actual uncompensable evils. Have I misunderstood?
Right, both (Carl and Jeff). Sorry to be (really!) sloppy. I'll make that change.
And it doesn't seem implausible that there really are actual uncompensable evils. Have I misunderstood?
No, I do intend to make the claim that there is no actual evil E such that there is no possible R meeting (i) □(R ⊃ E) and (ii) R & E is positive. Of course R might be a conjunction. I don't find that implausible. For any actual E, the disvalue of E is finite. So, for some finite conjunction of r's = R, (i) and (ii) will hold for R and E.
As frequently happens to me, I'm missing something. Suppose that for each actual evil E, there is a possible R that meets your conditions. Why is this sufficient to establish that there is no gratuitous evil in the actual world? Doesn't this require that for each actual evil, a positive response exist, not just be possible? If God prevents E under your conditions, he is preventing a possibly greater good, not an actually existing greater good.
Mike:
How exactly do you understand a gratuitous evil?
Suppose that J responds to E with R, but if E' happened in place of E, then J could have instead responded to E' with R', where E' is much smaller than E, and E'&R' is more valuable than E&R. In that case, would you count E as gratuitous? It seems that a case can be made that in such a case E is gratuitous, for while God could get the good E&R without E, God could have gotten a greater good, E'&R', without E.
Even if for each evil E in the actual world, it's possible that there is an R & E in the way suggested that makes E non-gratuitous, it doesn't follow that the various responses to the evils in the actual world are compossible. The argument doesn't rule out that there is an E such that the justifying R & E for it rules out the R for some other E.
If God prevents E under your conditions, he is preventing a possibly greater good, not an actually existing greater good.
I guess I don't know what it means to prevent an actual greater good. If the good is actual, then it wasn't prevented. The only goods that can be prevented are ones that are possible.
Even if for each evil E in the actual world, it's possible that there is an R & E in the way suggested that makes E non-gratuitous, it doesn't follow that the various responses to the evils in the actual world are compossible.
No, that's right. But why does it matter? All I intended to show is that there is no evil that is gratuitous. And in any case, I don't see any reason to deny the additional claim.
How exactly do you understand a gratuitous evil?Suppose that J responds to E with R, but if E' happened in place of E, then J could have instead responded to E' with R', where E' is much smaller than E, and E'&R' is more valuable than E&R. In that case, would you count E as gratuitous? It seems that a case can be made that in such a case E is gratuitous, for while God could get the good E&R without E, God could have gotten a greater good, E'&R', without E.
Alex, this is a nice question. The concept of gratuitous evil badly needs to be thought through more cautiously than it has been so far. There are nice complicating examples to consider. I'm using a standard view on which a non-gratuitous evil E is such that necessarily, God does not prevent E without preventing a greater good G (or permitting a greater evil, E').
Does animal suffering count as evil?
If so, your claim (1) about the actual world seems false.
I'm using a standard view on which a non-gratuitous evil E is such that necessarily, God does not prevent E without preventing a greater good G (or permitting a greater evil, E').
But doesn't this say that necessarily, if God prevents E, he prevents a greater good G that would have existed not just might have existed, had he allowed E? Your argument seems to show only the latter.
Does animal suffering count as evil? If so, your claim (1) about the actual world seems false.
Animal suffering is extensive but finite. So, whether there is a response R to animal suffering E such that R&E is positive is an empirical question. People can act in positive ways that are directed essentially to that suffering. Smith might commit his life to positive moral action on behalf of animals in commemoration of those particular sufferings. He could not do that had those animals not suffered. I don't mean that callously or blithely, since I completely agree that animals have suffered immensely. It would take a lot of positive action to produce an R such that R&E is positive. But I think it's possible.
I tend to prefer some definition like that an evil is gratuitous if an omniscient and omnipotent being would not be morally justified in permitting it.
The more standard definition you give just doesn't seem strong enough. Suppose that E is necessary for G, and G is a greater good than E is an evil, but were God to prevent E, a much, much greater good G' would result.
If we want to stick with the standard definition, what we want to say is more like this: E is necessary for getting some good at least as great as G.
But that's still not quite right. For suppose that there are two evils, E and E', which are equally bad, and E isn't necessary to getting G (or something at least as good), because E' will also suffice for G. That doesn't make E gratuitous, even though it is true that E is not necessary for G. So, we may need to tweak as follows: something at least as bad as E is necessary for getting something at least as good as G (or avoiding something at least as bad as E2).
I think there are still technical problems with this formulation, but it gets closer to the nub of the matter.
So, now, I'm worried that your argument only works on the flawed definition, but not if one improves it in this way.
But doesn't this say that necessarily, if God prevents E, he prevents a greater good G that would have existed not just might have existed, had he allowed E?
We're conflating views about improving the world with views about preventing evil. You seem to be suggesting either that (i) E is an instance of non-gratuitous evil iff. □(R ⊃ E) & □(E ⊃ R), or (ii) E is an instance of non-gratuitous evil iff. □(R ⊃ E) & E □-> R. But (i) is way too strong, since no evil entails a response to it, and certainly a response would render it non-gratuitous. And, by my lights, (ii) is also too strong, since it makes God a moral actualist. Suppose you and everyone else should be responding to evil E appropriately but refuse to do so. If (ii) is true it is evidence against God's existence that you and everyone else refuse to bring about the best, E & R, by responding to evil appropriately. The world is worse as a result of what you and others do, but why does that count against God's existence?
E is necessary for getting some good at least as great as G. But that's still not quite right. For suppose that there are two evils, E and E', which are equally bad, and E isn't necessary to getting G (or something at least as good), because E' will also suffice for G. That doesn't make E gratuitous, even though it is true that E is not necessary for G.
In this case (E v E') is necessary for G since □(G ⊃ (E v E')). Assuming it is also true that ~□(G ⊃ E) & ~□(G ⊃ E'), we get that it is permissible that God prevent E, permissible that he prevent E' but not permissible that he prevent both.
That's true--if you introduce disjunctive evils and disjunctive goods, you can handle some of these problems. But once you have disjunctive evils, (1) becomes a bit less plausible (because you've increased the amount of evil to quantify over).
Disjunctive evils are kind of weird, too.
I tend to think that your argument works for the terms as you have described them, but it only takes a slight semantic shift to circumvent your point.
As of now, you say that evil is gratuitous (or not evidence against God's existence) if it is impossible that said evil brings about a greater good. Sure, I'll grant that. And I'll grant that no actual evils have this property. Thus, no actual evils are evidence against God's existence.
However, on your account, the evils themselves become non-problematic only because there may be a good that makes them worthwhile. Thus, while the evil itself is not evidence against God's existence, the conjunction of the evil obtaining and the appropriate good not obtaining is now evidence against God's existence. E is no longer a problem, but (E ^ ~R) is now a problem. Is passing the buck in this fashion really a big improvement?
E is no longer a problem, but (E ^ ~R) is now a problem. Is passing the buck in this fashion really a big improvement?
I guess I don't follow you. Concerning "passing the buck", if anything, I'm trying not to pass the buck. The buck for there being no response R such that R&E is positive stays with us. The existence of evils for which there is no actual compensating good reflects poorly on us, not on God. For any evil E, at any time of occurrence, it is possible for some individual I (or some set of I's) to respond at any time to E in such a way that the response R is essentially a response to E and E&R is positive.
So what't the argument? If God existed, he would see to it that he responds to E instead of me? Why not argue that if there's really a God, he'd drive me home everyday, instead of making me do it.
Mike:
In your first claim, are you quantifying over all actual individuals or over all possible individuals (i.e., is the quantification over individuals outside or inside the scope of the possibility operator)?
In your first claim, are you quantifying over all actual individuals or over all possible individuals (i.e., is the quantification over individuals outside or inside the scope of the possibility operator)?
Alex, I worried about that. I did leave it inside the scope of the operator, and assumed that the relevant world was accessible from ours. That is, it might not be true that every individual who can contribute R actually exists. He might be a future person. On the other hand, I do not want to consider possible, non-actual individuals who are not future persons. So, I can't quite represent what I want to say with either the narrower or wider scope for the possibility operator. I've got to restrict the relevant worlds, or stipulate that future persons exist. Something along those lines. I guess I could argue as well (though it did not occur to me) that this is another reason why it matters that there exist sufficient numbers of human beings. In that case even possible, non-future persons are relevant to the argument.
Animal suffering is extensive but finite. So, whether there is a response R to animal suffering E such that R&E is positive is an empirical question. People can act in positive ways that are directed essentially to that suffering.
I can see that there are responses R that require that there be animal suffering (the sorts of responses you list). But my worry is that lots of the particular instances of animal suffering have no response R (e.g. Rowe's fawn caught in the fire).
Mike:
And I think you need more than logical possibility of responding. If it's logically possible for an existent person to respond in the relevant way but it's not nomically possible, then we can't hold that person responsible for the non-response. There is some world where I have the superpower to precipitate water out of the air in large quantities. In that world, I can respond to the initial moments of the fawn's being caught on fire by putting out the fire, and my response plus the evil is a net plus, because I've exercised virtue, and the fawn only suffered for a few moments. So it's logically possible that I be able to responod to the fawn's suffering. But that's not the relevant kind of possibility. One needs an actual causal possibility for the argument to work, right? And that makes it quite a bit harder to make (1) plausible.
I can see that there are responses R that require that there be animal suffering (the sorts of responses you list). But my worry is that lots of the particular instances of animal suffering have no response R (e.g. Rowe's fawn caught in the fire).
I'm not sure why you say that. There's an R for Rowe's fawn as well, since there is an R such that R is essentially a response to the suffering S of Rowe's fawn and R&S is positive.
There is some world where I have the superpower to precipitate water out of the air in large quantities. . . So it's logically possible that I be able to responod to the fawn's suffering. But that's not the relevant kind of possibility. One needs an actual causal possibility for the argument to work, right? And that makes it quite a bit harder to make (1) plausible.
Superpowers are not the sorts of things that I had in mind. They're not necessary to most moral responses R that outweigh evils. What I have in mind is something people without such powers can do. Here's the example I gave to Tim above,
People can act in positive ways that are directed essentially to that suffering. Smith might commit his life to positive moral action on behalf of animals in commemoration of those particular sufferings. He could not do that had those animals not suffered. . .
Smith might live morally in commemoration of S, where S is the suffering of animals that have undergone certain sorts of abuse, for instance. Had S not occurred, it would be impossible for Smith to act in just the way he does. Such lives can be extremely valuable and, I'm urging, can generate a response R such that R&E is positive.
There's an R for Rowe's fawn as well, since there is an R such that R is essentially a response to the suffering S of Rowe's fawn and R&S is positive.
There is? Please describe it. (A serious, not a rhetorical request.) I thought the main upshot of Rowe's fawn was that no people observed or learned about its suffering, so that it couldn't be the occasion for the possibility of the sort of praiseworthy responses R you seem to have in mind.
Rowe's fawn was that no people observed or learned about its suffering, so that it couldn't be the occasion for the possibility of the sort of praiseworthy responses R you seem to have in mind.
This is tricky, since Rowe's fawn is not an actual fawn. It's a hypothetical fawn. But it suggests that there are actual fawns more or less like the one he describes. Concerning those actual fawns, whether or not you actually see one, you can make them the object of your response, since you can refer to them in the description of what you do. I can, for instance, refer to every fawn that died in forest fires last year and commit myself to acting in ways that commemorate that suffering and preventing further unnecessary suffering. I needn't have ever seen such a fawn.
It seems to me this argument requires that God not have providentially useful foreknowledge.
For suppose God has such knowledge. Then God knows for any E whether E would in fact constitute part of an overall better state of affairs.
So God is not going to just allow a fawm to burn up on the off chance that someone responds positively to it. From God's perspective, there is no chance involved.
I think this illustrates a theological advantage of open theism, but I am not sure that was waht Mike intended.
So God is not going to just allow a fawm to burn up on the off chance that someone responds positively to it. From God's perspective, there is no chance involved.
Gordon, I don't assume open theism or believe open theism is true. The possibility of respondng morally to the suffering fawn is always open. Since that's so, the gratuity of the evil is something that we can eliminate. Should we decide to do so, of course, the evil would not be gratuitous. So the existence of gratuitous evil does not reflect badly on God, it reflects badly on us. We've chosen not to eliminate gratuitous evil, but just to complain about it. The existence of gratuitous evil would reflect badly on God if it were such that only God could eliminate it. But that's not so.
I guess I am just missing something. My obligation to save Joe from drowning is not obliterated by Sally also being able to save him. Of course if I know Sally is saving him, I don't have to do anything. If I know Sally is not saving him, then I am obliged, as well as if I am not sure what Sally is doing.
How is the situation with God different?
My obligation to save Joe from drowning is not obliterated by Sally also being able to save him.
I don't follow the example. Are you suggesting that every time you fail to fulfill an obligation or do something disvaluable, God ought to make up for it? If that's true, then there is overwhelming evidence that God does not exist. But I see no reason to believe it is true.
Its not a matter of God "making it up for me" its a matter of a certain bad state of affairs that God could have prevented but did not, knowing that no one else will prevent it.
Actually the case is worse than that, since plausibly there are cases in which no human being really has an obligation to prevent something bad from happening because they don't know it is going on, as in Rowe's fawn.
What is required is some reason for God not intevening when these bad things actually do occur. I don't think just saying it is possible for some creature to jump in gets God off the hook.
And of course there are cases in which it looks to be physically impossible for some creature to prevent the suffering. no finite creature I know of can prevent the suffering that occured right after the Haiti earthquake.
Of course there are candidates to fit the bill. Soul making theodicy is one. Maybe it works, maybe not but at least it tries to deal the question of why God allows actual evil (natural or moral).
Its not a matter of God "making it up for me" its a matter of a certain bad state of affairs that God could have prevented but did not, knowing that no one else will prevent it.
Again, this misunderstands the argument. The argument does not claim that anyone is required to prevent suffering. I claim that for any instance of evil E, there is a response R such that R is essentially a response to E and R&E is positive. For any evil E that has ever occurred at any time in the history of the world, it is possible to respond to E with R in the way noted. The epistemological worries are unproblematic, since I can refer to all of the unknown evils and direct my actions and intentions accordingly. It is true now that such responses are possible, and that they remain possible. So, the observation that there are evils E for which there is no R is not evidence that God does not exist.
Mike,
Sorry my comment is so delayed. I'd forgotten that I'd weighed in on this debate. So this is a response to your 2/26 (1:43 PM) comment.
I see your point regarding my accusation of passing the buck. The onus has shifted from E's existence to R's non-existence, but you claim that R is entirely on us. But if there are possible individuals who would respond with R (which is true by hypothesis), why didn't God create the world in which they DID respond as such? Doesn't this now just reduce to the free will defense?
Mike,
Your argument seems to be much stronger than I at first thought. One question: is it correct that you are not committed to either '(E & possible-R) has greater value than (not-E & not-R)' or '(E & the probability-of-R) has greater value than (not-E & not-R)'?
One question: is it correct that you are not committed to either '(E & possible-R) has greater value than (not-E & not-R)' or '(E & the probability-of-R) has greater value than (not-E & not-R)'?
David,
No, right, I don't either of those can be true, so I hope I'm not committed to them.
But if there are possible individuals who would respond with R (which is true by hypothesis), why didn't God create the world in which they DID respond as such? Doesn't this now just reduce to the free will defense?
James,
FWD does not include anything that I say here. But I would supplement FWD with this account of why free will is valuable. In fact, that was the point of developing the account, so that's a nice question.
One worry - what if the argument is reversed to look at omissions of evil? It seems highly plausible to me that there are some evils which satisfy [](R --> E) and (R&E) is positive that nonetheless haven't occurred. Unless I'm missing something, wouldn't these omissions be gratuitous?
Unless I'm missing something, wouldn't these omissions be gratuitous?
I'm not sure how non-occurring evil could be a gratuitous evil. I'm even less sure how it could count as a reason against theism that there are various sorts of evils that have not occurred. Maybe I'm misunderstanding you.
Mike:
I'm pretty sure the misunderstanding is on my part, but here goes anyway:
My idea was that privations of goods are the same as evils. God shouldn't prevent a good G unless doing so incurs a greater evil E (or would prevent a greater Good).
If so, then your fairly broad scope of what can make an evil non-gratuitous also raise questions as to why God seems to have missed so many opportunities for outweighing responses. It would seem pretty easy to find an example of an evil that seems likely to have satisfied your criteria but didn't occur (my death at the age of five, say). If so, why did God prevent the opportunities for these goods?
I'm pretty sure both steps in that argument are badly wrong, but I thought I'd clarify what I was driving at on the off-chance it's on to something.