Evil and Compensation

Dr. Pruss,

I think your point of a "generous compensation" is advantageous for the theist and the problem of evil for those who will enjoy the benefits of a properous generous compensation. However, how does your argument factor in the evils experienced by those who will not experience this kind of compensation? It seems that in the case of these unfortunate ones, having no "career" would be infinitely greater than any career that is eternally detrimental and generously compensated.

Mr. Moore,

I take it you're referring to the souls in hell. Well, the argument from evil augmented as you suggest is not going to be an argument against the existence of God, but against specifically Christian doctrine, namely the doctrine of eternal damnation.

Well, if hell can be justified, and if the sufferings in hell are infinite (or finite but very, very big), then the generous compensation account can still be given for the souls who go to hell. In compensation for not saving Jones from cancer, God decreases his suffering in hell for eternity, thereby giving him an infinite (or finite but very, very big) recompense. I am not claiming this actually happens, but there seems to be nothing absurd about the idea.

I suppose the really hard question here is about two forms of the problem of evil:
(a) the problem of hell; and
(b) the problem of the badness of doing wrong.
On reflection, the compensation account does nothing to help with either of these varieties. In regard to (a), one needs to say exactly what one would have said without the compensation account. (I take the line that it is better to exist in eternal torment than not to exist at all, and that justice is a good to the person being punished.)

But I do want to say something about (b). By "problem of the badness of doing wrong", I mean the question of why God, in considering the good of Jones, allows Jones to sin. (This is to be distinguished from the question of why God, in considering the good of Jones' victim, allows Jones to sin. Compensation does help with the latter question, since God can compensate Jones' victim.) It just does not feel right to say that God compensates Jones for having allowed Jones to sin.

Fortunately, the problem of the badness of doing wrong is, I think, fairly well handled by free will considerations.

Alexander thank you for an interesting post.

I'm not sure about three things

1. You quite rightly accept that the explanation will only satisfy the atheist if they are a consequentialist. I don't think the compensation move helps in this case to get the deontologist on board. If you look at the example you give there are two important elements present that make the compensation plausible: 1. The over compensation and 2. The time pressure the agent is under when they decide to compensate rather than warn. Indeed I would think that 2. is what does much of the work underwriting the compensation position.

But this second element simply isn't present with God if God is defined in the traditional omni-God fashion. In other words it looks like compensation only lowers the bar if the bar has already been lowered by limited abilities...
Now the argument might be revivable for a certain subset of evils, namely those that are logically necessary to bring about some greater good since on many accounts even God is limited by logical necessity although a deontologist may still reject this broadly consequentialist justification.

2. Secondly the compensation argument as your example shows depends on you not being the primary wrong doer, but instead someone who can avert the wrong but does not. But this is arguable in the case of the classic God, since as the creator arguably everything that occurs is God's fault (since they set up the universe that way).

3. Another deontological worry perhaps is that the very notion of compensation implies that the agent compensating has done wrong that they are redressing. What this seems to imply is that at different time slices God is not behaving in an omni-benevolent fashion. In other words on a whole life & afterlife view they are benevolent towards X but at particular time points when great evils occur to them they are not benevolent.

If you want some more reading on this topic send me an email, this is basically what my master's thesis focused on.
Cheers
David Hunter

Alex
Seems like God in your scenario would not get very high marks as a manager. When I was in business if I allowed instances of preventable harm to occur there would have been a high cost to pay which would have increased the cost of my product and lowered profit. I would have been fired had I knowingly and freely allowed this to occur. On utilitarian grounds alone, 'compensation for harm' is not a very good concept to operate on if the manager does not try to eliminate the causes of harm so as to lower the need for compensation. If God is not a utilitarian, then there are other sound reasons for allowing preventatable harm to occur that would eliminate the need for compensation to justify the harm.

David:

Thanks for these helpful comments. You are right that the compensation depends on God not being the doer of the evil. But that God is not the doer of the evil is something we are always going to need to say in any response to the problem of evil.

Deontologists generally need distinctions like doing/allowing or intending/foreseeing. A deontology without these distinctions is going to work very poorly. These distinctions must be brought to bear on the problem of evil. This is easiest to do in the case of those evils that are the foreseen consequence of indeterministic processes (such as free will or quantum stuff) where God does not will a particular outcome.

I don't think compensation entails a wrongdoing, though the word sometimes does. But we do talk of "worker's compensation". Anyway, in the example of the person who doesn't have time to stop the conman but who plans on compensating, the person has not done anything wrong. The reason is that whether a preventing of an evil is morally required depends on various things. One of the things it depends on is whether one is planning and able in some way to "undo" the evil, not literally but by bestowing a much greater good later. It is one thing to sit by idly while an evil is happening, and another wait while having a more than satisfactory response planned. What plans one has in mind certainly does affect the moral nature of what one is doing.

You're right that the time pressure in the divine case has to be replaced by something like the idea of goods that logically cannot be made to occur without God's also permitting the evils. But this does not make the notion of compensation useless, since it lowers the bar on how great these goods would have to be. Thus one might think that the Holocaust is too big a price to pay for the relevant kind of significant free will and the sorts of virtuous actions (e.g., among rescuers of Jews) that it made possible. However, when one adds in generous compensation, the "too big a price to pay" issue becomes less pressing.

John:

I think the poor management account only applies in cases where the compensation costs the company something. In a case where the compensation does not cost the company anything, there is no poor management issue.

Alex
By definition 'compensation' is a cost. There is no instance where it does not costing the company something. To state otherwise is simply ad hoc.
John

John:

Compensation is a gain to the person compensated. It need not be a loss to the person compensating. Here's an example. Software company sells a game for $50. They never sell it for less. I would never pay that much for a game. To compensate me for something or other, the company gives me one free personal license. Since I would never have bought the game, they lose nothing by giving me a free license. But I gain.

Alex
You are not being compensated for not purchasing the game by the company giving you the game for $0.00. You have lost nothing that the company needs to compensate you for. Personally they do not care if YOU buy the game as long as enough other people buy the game to make it more profitable for them to sell it at 50.00. If no one buys the game at 50.00 then they will lower the price. This lower price is not compensation.

If they produce 100 games and can sell all 100 games for 50.00 then giving you a game for free does cost them something. But, there is no incentive to give you a free game. What are they compensating you for? What have you lost that they need to compensate you for?

You are not really addressing the issue I originally raised. If a manager can produce x at a cost less then she is presently doing, then she is not a very good manager. If God can create a world with less evil that would reduce the need for future compensation and does not do so, He is not a very good manager of the world he creates.
John

John:

I am assuming that the free license is given to me in compensation for something. (For instance, maybe the company owner's son was rude to me that morning, and the owner decided to apologize to me by giving me that free license.) I didn't say they are giving me a copy of the game--that would, indeed, carry a minor cost--but only a license. A license is just a permission to copy and use.

The management analogy, I think, only applies when the compensation presents a real cost to the company. But when the compensation costs the company nothing, I see nothing wrong with the management. I suppose one might have an esthetic objection--one might think it's messy--but I am not sure how strong that objection is.

Hope this system of quote indicating works, here goes.

>>You are right that the compensation depends on God not being the doer of the evil. But >>that God is not the doer of the evil is something we are always going to need to say in >>any response to the problem of evil.

I’m not so sure we have to say this in every response to the problem of evil. Depends on what we mean by evil but supposing we are meaning causing suffering then it seems in some cases that this can be perfectly acceptable to do this. So take the dentist for example, they cause suffering but this evil is okay because 1. On the whole the action is good for the patient and 2. The patient consents.

If any of the evils resemble this pattern then it would seem fine to me that God does them rather than just allows them. The consent is going to be the sticky part, I believe (but could be wrong) Peter Forrest argues that in certain circumstance we can assume people would have consented and that would be enough. I’m skeptical of that move though.

In any case, fair point lets set that objection aside.

>>Deontologists generally need distinctions like doing/allowing or intending/foreseeing. >>A deontology without these distinctions is going to work very poorly. These >>distinctions must be brought to bear on the problem of evil. This is easiest to do in the >>case of those evils that are the foreseen consequence of indeterministic processes (such >>as free will or quantum stuff) where God does not will a particular outcome.

Agreed, but I don’t think any of those distinctions will wash all that well when we are considering the actions of something which is omnipotent and omniscient. The most plausible examples of these distinctions usually have an agent with clearly limited powers and/or knowledge. The best counter examples (for example James Rachel Kill and Letting Die examples) the agent has full knowledge and the power to intervene at no real cost to themselves.

>>Anyway, in the example of the person who doesn't have time to stop the conman but >>who plans on compensating, the person has not done anything wrong.

Not what my intuitions say, particularly if they didn’t compensate. In any case imagine that our conned person and the possible preventer met up and the conned person found out that the possible preventer had known about the con but hadn’t stopped it. While they may be grateful that the compensation occurred, I think they might also be rightfully annoyed that the con wasn’t prevented in the first place. I think part of their rightful annoyance is simply that the possible preventer decided for them and didn’t give them a chance to decide for themselves. Nor do I think the rightfulness of their annoyance decreases as the compensation gets better.

>>The reason is that whether a preventing of an evil is morally required depends on >>various things. One of the things it depends on is whether one is planning and able in >>some way to "undo" the evil, not literally but by bestowing a much greater good later. >>It is one thing to sit by idly while an evil is happening, and another wait while having a >>more than satisfactory response planned. What plans one has in mind certainly does >>affect the moral nature of what one is doing.

Fair point, but I’m still not convinced, in fact I’m becoming increasingly convinced that what is bothering me is the cavalier fashion an autonomous individual is being treated, simply being compensated doesn’t quite make up for it to me, since this effectively treats them as if they can be bought off and have no say in the matter…

There are two problems with this:
1. There may be evils that are cannot be compensated for. I believe Marilyn McCord Adams discusses this in Horrendous Evils and the Goodness of God. She suggests that things like rape and other events which may literally destroy the person who suffers them cannot be compensated for. I can’t recall her resolution of this, except that I found it profoundly uncompelling.

2. Even if there is compensation the sufferer may refuse to accept it, you can imagine Richard Dawkins doing this… But a better example is one particular case where someone asked for euthanasia after a serious accident which left them disabled. They were refused, and now are happy to say they are happy to be alive (so refusing benefited them) however they still feel that their wishes out to have been respected and that they were disrespected as an autonomous individual. I am sympathetic to their complaint, regardless that they are better off because of it there is something a deontologist ought to be bothered by here.

>>But this does not make the notion of compensation useless, since it lowers the bar on >>how great these goods would have to be.
I certainly agree that the notion of compensation isn’t useless, it is just limited in regards to the type of evils it can explain to a deontologist.

Finally a standard problem for theodicies that rely on heaven, what my thesis supervisor used to call no eschatology without revealed eschatology… In other words appealing to the goodness of heaven seems a bit like cheating unless there is a darn good reason why this goodness is hidden (and for that matter not present here and now)

David (if I may),

Ah, yes, autonomy, that beast. Personally, I don't really have much use for the concept of autonomy. The space that others fill with the concept of autonomy I fill mostly with the concept of vocation. I personally find it amazing that some people think we typically have the right to choose, rather than discern, what job to take or whether to get married or the like. (On what grounds does one make such a choice?)

I think something like a respect for autonomy is important in interactions between humans, because we often do not know what vocation God has given to someone else, and because the task of one is not the task of the other. But of course that would not apply in the case of God.

God asking someone what they want seems very strange within Judaism and Christianity. The only instance I can think of is the case of Solomon being asked what he wanted, and that was to some extent a test, I think.

The following conditional seems extremely plausible: If there exists God, then creatures do not have much if any autonomy over and against God. After all, God has authority over creatures--he can command them.

The distinction between foreseeing and intending is not so much to be grounded in particular examples, as in the metaphysics of the will. :-)

Interesting.

How does this apply (if at all) to the problem of (non-human) animal suffering?

You might say, "No problem-- All dogs go to heaven!" But I'm not so sure. Leaving aside theological problems for the moment, I wonder whether this works for all animals.

The issues get very tricky here, but I suspect that there is a "sweet spot" on the phylogenetic scale where the animal is able to suffer, but doesn't have the resources (actual or potential) of memory, personality traits, etc., where it would make much sense to speak of anything resembling personal identity--even though there can be identity of the organism--such that the same subject can be later compensated for the evil it earlier suffers.

Well, I think that personal identity for us just is the same as our identity as organisms, and has little to do with the continuity of memory, personality traits, etc. :-)

But even if it turned out that we weren't animals, and that personal identity for us didn't go along with the identity of the organism, it seems that if you took the animals at the level of that "sweet spot", then you'd want to say that it is the animal as such that feels pain. But then the identity of the subject of the pain will just be the same as the identity of the animal, and the problem of identity will also not be present.

(There is, of course, the issue that as a matter of fact I don't think animals get compensated in an afterlife. But they may get compensated in various ways in this life. Who knows?)

Alex: You write: "I am assuming that the free license is given to me in compensation for something. (For instance, maybe the company owner's son was rude to me that morning, and the owner decided to apologize to me by giving me that free license.) I didn't say they are giving me a copy of the game--that would, indeed, carry a minor cost--but only a license. A license is just a permission to copy and use."

This does not help your argument. In fact it supports mine. You are granting that compensation is a form of making right that which was wrong. How can a perfectly good being be involved in doing something wrong which would require Him to compensate someone for a wrong that He is ultimately responsible for? Now even if you maintain that He is not the one doing wrong, your use of compensation and who gives the compnesation indicates that He is responsible to correct that wrong in some fashion that makes up for what was lost by giving something of equal or greater value to that person who was wronged. The ownder did not do the wrong to you (in your example) but as the owner he is responsible for the action of those who work for him and is therefore resposnible for compendsating you for the harm done by his agents. When I was a Director of Operations I was responsible for the business related actions of those that reported to me. If they performed a wrong action, and for purposes of this discussion, an action that potentially increased the cost of doing business, I was ultimately responsible for making sure that corrections (compensations) were made. I could not do a 'Pontius Pilate' and walk away claiming that I had no responsibility.

The management/economic model works very well for understanding God's relationhip to His creation. He too cannot do a 'Pontius Pilate.' God is just like the business person who starts up a new business. If the business person is doing her job correctly she will set up the business so as to produce the goods/services at the lowest possible cost (with the greatest possible benefits to her and the other stakeholders). In your initial example this was not done. The 'business' was set up so as to incure increased costs and this is not good management nor good economics.

By the way, the arguement I am using works for punishments also.

Tim, funnily enough from my point of view animals are much easier to compensate than humans (if we are taking about lower level animals). This is because there is no need to take heed of their autonomy in the same way we might think that we ought to in regards to humans.

Alex, I take your point, and can see how that is perfectly internally consistent in regards to your own beliefs. But given that your argument was initially hoping to get Deontologists on your side, I'm not convinced that denying the meaningfulness of autonomy is really the best stratagem. Surely they will argue that treating humans as autonomous agents is simply what behaving towards them appropriately entails.

However I share your wariness to infer from how humans ought to behave towards each other to how God ought to behave towards humans.

Cheers
David

David:

Well, not all deontologists think much of autonomy. I'm a deontologist and I don't think much of autonomy. :-)

Moreover, it may well be that the intuitions about autonomy only apply between equals, and hence would not apply between God and human beings, particularly if God is the Good Itself, and hence all decisions in some way aim at union with God.

Do you really think there are many people who think that if there were a God, he wouldn't have an innate right to command us? I know that my colleague Mark Murphy defends a view kind of like this in his excellent book on divine authority, but he is very much in the minority.

David:

I agree that issues of having to respect autonomy won't arise, so that's one way in which compensating lower-level animals would be easier.

Cards on the table: I'm a Partfitian on personal identity, and maybe if you're not one (as Alex isn't), the worry I was trying to express above is hard to motivate. But here is the idea, anyway: for the sake of argument, let's presume that a goldfish can feel pain (this isn't at all clear, but leave that aside), has no long-term memory, and doesn't have much of anything we'd want to call personality traits. Let's imagine that I inflict great suffering on some goldfish. Can I compensate that goldfish for that evil by treating it quite well 6 months later? I have serious doubts about that. If I treat my dog Fido badly, then--even though Fido isn't a person--I think it might make sense to talk about making up for my past ill-treatment by doing right by him later.

FWIW, if you don't bring in an animal afterlife, as Alex doesn't, then there seems to be loads of animal suffering that isn't compensated for in this life, and to which many of the usual theodicies (e.g., free will, soul-building) don't apply.

Tim:

If mental states don't supervene on physical states, then it may be that God compensates animals by intensifying some of their ordinary animal joys without affecting them physically.

Alex: for this to work, you'd have to presume, not only that mental states don't supervene (and aren't identical to) physical states, but also that mental states don't causally affect physical states. Otherwise, when God compensates animals by intensifying their joys, we'd expect to see the usual physiological and behavorial effects of having such experiences (which we don't).

I guess that God could also intervene to prevent those mental states from having their usual effects, but (IMHO) invoking constant undetectable local miracles would be a desperate maneuver.

Tim, I am not a Parfitian on personal identity, but I still see your point. I think you can motivate it in a different fashion even for someone who isn't a Parfitian to much the same effect. This was the appeal I was making to the Horrendous Evils point, we may think that some evils are so horrific that they actually change the person who is present, perhaps destroy their identity. If this is the case then your point seems to stand, since these horrendous evils would be uncompensatable.

Tim:

Would we actually notice if God miraculously boosted a dog's enjoyment of a drink of water, even if this had the same effects as enjoyment usually does? The dog would just seem extra happy that day--such variation does happen.

David
You write "However I share your wariness to infer from how humans ought to behave towards each other to how God ought to behave towards humans."

The problem I have with this is that if we can ascertain what a good moral agent will do and this is determined by how she treats others as a result of guiding her actions by morally acceptable principles how can we even begin to assert that God is good if we cannot infer from what we know to how He should relate to other moral agents? We need a compass to direct us in how to evaluate the actions of others and if God is acting towards others then our knowledge of what constitutes a warranted moral framework needs to be that compass.

John Alexander

Alex:

Sure, God could miraculously make Fido feel 2n hedons of pleasure when drinking water instead of merely n hedons, and we may not notice the difference. (I find the notion of God constantly doing this sort of thing really strange, but let's leave that aside.) But then we'd still have excellent empirical grounds for thinking that lots of animals have (often brief) lives containing truly horrific suffering with no comparable compensating joys. God's going around surreptitiously squinching up the hedons here and there in ways we don't notice wouldn't substantially alter the situation.

Tim:

Remember, though, that some of the aspects of suffering that are there for humans--wondering when it well end, etc.--may not be present for animals.

And just as God can double the pleasure, he can halve the pain.

Alex,

I don't know whether this point is made in the comments above. But there is an obvious objection here. The agnostic's complaint against C* is too weak. C* is the infinite life/afterlife whose goodness washes out any finite amount of pain.
The objection is that there is yet another career relevant here, C+. In C+, my finite life includes no suffering and I also get an afterlife of infinite positive welfare. Obviously C+ is better than C*, even if the Cantorian cannot make the math reflect that. So much the worse for Cantorian representations. There are all sorts of nonstandard treatments of the infinite on which addition, subtraction, etc. are well-behaved. Choose a nonstandard treatment for a more accurate representation of the fact that C+ is better than C*.
The agnostic argument then goes that a perfect being would have actualized C+ (only a sub-perfect being would actualize C*). That's essentially the same complaint he started with. The addition of infinitely good afterlives doesn't help the perfect being theist.
There are incidentally other excellent reasons to go nonstandard in your treatment of the infinite. Suppose you stick with Cantor. In that case a perfect being would be justified in giving each agent a .0001 chance of having your career C* and a .9999 chance of having a finite career C- of nothing but unbearable (though finite) pain. The value of the mixed strategy M is infinite, M = .0001(C*) + .9999(C-). And the value of M equals the value of P, the pure strategy of giving every agent the infinitely positive career C*. So a perfect being is justified in using P only if he is justified in using M: they have exactly the same value. Obviously, something has gone wrong. What's gone wrong is the standard Cantorian representation.

Mike:

I am not sure I understand your objection. Principle (I) is fully general. In particular, we can substitute C+ for C, and (I) will yield a career C+* that contains finite evils, and whose earthly extent is just as we observe, but which is infinitely better than C+.

The reason for this is that even if we imagine a person having a future with infinite bliss, God could make that bliss infinitely bigger, for instance by doubling the amount of bliss at any given time.

Alex,

You say that (I) significantly alleviates the problem of evil. But it doesn't. Here's (I).

(I) Let x be an actual person who suffers the set E of finite evils in his earthly life. Let C be any possible career of x. Then there is a possible career C* such that: (a) C* includes x's suffering the evils from E, (b) for aught that we know, C* is x's actual career, (c) it is better to exist with career C* than not to exist at all, and (d) x would be better off by an infinite amount in having C* instead of C.

You let C be any career of x that includes some amount of finite evil E. You suggest that the problem of evil is mitigated by the fact that for all we know C is a proper subset of C*, where C* includes an afterlife of infinite positive value. But that does not mitigate the problem for the atheist. What the atheist wants to know is why God did not actualize C+ instead. Recall that C+ includes NO FINITE EVIL at all. C+ is a life of no suffering PLUS an afterlife of infinite positive value. C+ is better than C*.
Now I'm guessing you'll say: "oh, but for all we know C* has double the value we thought it had and so is even better than C+". But that's not true. Here's why. Let N be any infinite value you decide to include in C*. There is still the possiblity that C+ has the infinite value N, too, PLUS a finite life of no suffering at all. So no amount of positive infinite value in C* will mitigate the problem of evil.

Dear Mike:

Here's what I think is happening.
(I) basically claims: For any career without finite evil, there is a better career with finite evil. You claim: (J) For any career with finite evil, there is a better career without finite evil.

So far this looks like a tie.

But now consider this. It is highly plausible that there is some career which the atheist would agree that a perfectly good God could actualize.

Fix any such career. Call it C, C+ or whatever.

But if we grant the principle that if a perfectly good God could actualize C1, and C2 is better than C1, then God could actualize C2, we conclude that God could actualize a career that contains finite evil. And even if we do not grant that principle (I am inclined to think we shouldn't), we get the alleviation argument.

And even if we do not grant that principle (I am inclined to think we shouldn't), we get the alleviation argument.

Alex, how? I can't see how you get it even granting the principle.
Anyway, I thought you were making this interesting claim: Just given the fact that a career with evil can be included in an infinitely valuable career we have some alleviation for the POE. That's interesting, but false, for reasons given above.
Now I agree that there is a world that an atheist would grant God could actualize. But it would be a world that meets the standard conditions on permissible evil: viz., that God would not actualize any world that included a single instance of gratuitious evil. Let w be such a world. Now could God create a world with even more evil than w? I guess he could: viz., a world w' that contains more evil than w but no gratuitious evil AND where w' is as good as w.
But all of that is fairly obvious. I got the impression that you were saying that God could actualize a world with gratuitous evil provided that he put "lots of icing" on the world, so to speak. That's what I'm denying, for any amount of icing God might add. Am I following you?

Mike:

Well, given the principle, I think the conclusion follows fairly uncontroversially. Let w0 be the actual world. Let w1 be be a world with no evil that the atheist agrees is good enough that God would create. Then, by (I) (generalized to the case of multiple sufferers) there is a world w2 which is just like w0 in respect of the earthly life of creatures but which has so much icing that w2 is better than w1. Since w1 is good enough for God to actualize, so is w2 by the principle. But for aught that we know, we are in w2. Hence, for aught that we know, we are in a world that it is good enough for God to create.

Without the principle, we only get alleviation. And we get it in the following way. Suppose that there is a good G that logically requires an evil E. For instance, maybe some esthetic good G could only exist if E exists. (If we think of the world as a drama, then this is very plausible.) But this isn't enough to secure the non-gratuitousness of E, since G just might not be good enough to be worth paying the price E for. However, it may be that given compensation, it will be OK to choose G at the price E, while without compensation it would not be OK.

For instance, suppose that Jones' having significant free will, given some background assumptions, requires that God allow Jones to hurt Smith. Now it might be that the value of significant free will is insufficient to justify allowing Smith to be hurt. However, if Smith can be sufficiently compensated for the hurt, then permitting Smith to be hurt in order to allow Jones to have free will may well be permissible.

Let w0 be the actual world. Let w1 be be a world with no evil that the atheist agrees is good enough that God would create. Then, by (I) (generalized to the case of multiple sufferers) there is a world w2 which is just like w0 in respect of the earthly life of creatures but which has so much icing that w2 is better than w1. Since w1 is good enough for God to actualize, so is w2 by the principle. But for aught that we know, we are in w2. Hence, for aught that we know, we are in a world that it is good enough for God to create.

Alex, thanks for this. Here's where I think the argument goes wrong. I agree entirely that there is some world w1 that the atheist would say God might actualize. By stipulation w1 includes no evil and (let me stipulate) w1 has overall value V.
Now you want to argue that for any world w2 such that w2 has value V+n > V, the atheist is committed to saying that God could actualize w2 as well. After all, w2 is better overall than w1.
But this is mistaken, I'm sure. The world w2 might have a thousand times more value than w1, but if w2 contains any gratuitous evil, then God could not actualize it. The atheist's objection is not that w2 is not good enough. His objection is that it has evil that God could prevent without losing a greater good or producing a worse evil.

On the different point you say,

For instance, suppose that Jones' having significant free will, given some background assumptions, requires that God allow Jones to hurt Smith. Now it might be that the value of significant free will is insufficient to justify allowing Smith to be hurt. However, if Smith can be sufficiently compensated for the hurt, then permitting Smith to be hurt in order to allow Jones to have free will may well be permissible.

How is there alleviation here? There are three worlds. In w1 Smith harms Jones and there is no added payoff to Jones. In w2 Smith harms Jones and there is an added payoff to Jones. In w3 Smith is prevented from harming Jones and there is an added payoff to Jones. In terms of value, given the stipulation that the harm outweighs the freedom, we have w1 worse than w2 worse than w3.

Mike:

"The atheist's objection is not that w2 is not good enough." Fair enough. This is tantamount to a denial of the consequentialistic principle that it is always permissible to do better than something that is permissible.

"we have w1 worse than w2 worse than w3": Here things are more complicated due to incommensurability. I believe that goods of different kinds are incommensurable. The harm to Smith is probably incommensurable with the value of free will. In the presence of incommensurability, "better than" must be understood in terms of domination.

I did not say that the harm outweighs value of free will. Rather, I claimed that free will did not outweigh the harm. This is a different claim in the presence of incommensurability (and even without incommensurability).

Now, it may be that w2 dominates w1, because the added payoff to Jones can be very large and chosen to be commensurable with the evil that Jones suffers (e.g., pleasure is commensurable with pain, perhaps). However, w3 does not dominate w2, because there are instances of goods in w2 that are not dominated by goods in w3, namely free choices with the possibility of harming others. (I am assuming that God ensures there is no evil in w3, rather than this simply happening by chance because people choose not to do wrong.)

"The atheist's objection is not that w2 is not good enough." Fair enough. This is tantamount to a denial of the consequentialistic principle that it is always permissible to do better than something that is permissible.

I'm not denying the consequentalist principle. At least not here. Let w1 be a world with no gratuitous evil. Let w1 be just like w except for a "few more dancing girls", to coin Plantinga. I don't deny that there are always worlds like w1 that are better. I mean so long as we do not add gratuitous evil to w1 along with the goods.

But this is puzzling to me.

Now, it may be that w2 dominates w1, because the added payoff to Jones can be very large and chosen to be commensurable with the evil that Jones suffers (e.g., pleasure is commensurable with pain, perhaps). However, w3 does not dominate w2, because there are instances of goods in w2 that are not dominated by goods in w3, namely free choices with the possibility of harming others.

Suppose we stipulate such incommensurability worries. Now you said that w1 with the [freedom and harm] was such that "the freedom does not outweigh the harm" right? Let w* be just like w1 except without the [freedom or harm]. Obviously, w* is at least as good as w1 (I say obviously, because we have removed what is, by hypothesis, an on-balance bad or neutral state of affairs, viz. the freedom plus harm). Now add the same amount of good G to w* that you added to w2. Now w3 just is w* plus G. And w2 is just w1 plus G. Since w* is at least as good as w1, w3 is as least as good as w2. So incommensurabity does not undermine the argument.

I did assume initially that w* would be better than w1. But that is only because, if the state of affairs, [freedom + harm] is neutral, then there is no problem to begin with. The discussion collapses. There is no reason to add the goods of w2, since the freedom is worth the harm in w1.

Mike:

On incommensurability, I deny that w* is at least as good as w1. All that follows is that w* is no worse than w1. But "no worse than" is not the same as "at least as good as" given incommensurability.

Mike:

Another thing:
If the consequentialist principle is true, then God can create a world with gratuitous evils.

For take a world w1 that God could create, and suppose w1 includes at least one human. Add to the world two items, first a minor pinprick to one of the individuals in w1, and second a very large extra dose of (good) pleasure to that same individual, right after the pinprick, and call the resulting world w2. The relevant individual is better off in getting the tiny amount of pain plus a large dose of pleasure than in getting neither. Hence, w2 is better for that individual than w1 is. But w2 is at least as good for everybody else. Assuming that how good a world is is determined by how good it is for the individuals within it, we conclude that w2 is better than w1.

If the consequentialist principle is true, then w2 is permissible if w1 is. But w2 includes a gratuitous pinprick. Ergo, etc.

Alex,

I think I missed something important: I don't understand the domination. Could you explicate the difference beween "A outweighs B, and they are commensurale" and "A dominates B, and they are incommensurable"?

Yesterday, I read (in Peter Suber's article Infinite Reflections) that there is a notion of "incommensurable quantities" in math. I don't know what it consists in. Is that notion relevant for your discussion with Mike?

I suppose it is too late to jump in here again, but I do have a comment on your reply to Mike.

I grant that W2 is better then W1, but W3, a world that has great pleaure and no pinprick, would be better then W2. Assuming that God can create W3, then gratuitous evil is not justified. Now, God can create any possible world He chooses (assuming He is free) so if he creates W2 He is creating a less then better world.

If the pinprick is a necessary condition for the greater pleasure, then it is not gratuitous.

Vlastimil:

I take it that there are many different kinds of goods, and a thing that we call overall "good" may instantiate multiple kinds of good for different individuals. Thus, Jones' feeding of the hungry Smith instantiates at least the following goods:
- Smith's being nourished
- Jones' cooperating with Smith in a good activity
- Jones' acting virtuously

G2 dominates a G1 provided that for every individual x, and every kind of good K, either K is inapplicable to both G1 and G2, or G2 is at least as good for x in respect of K than G1 is. The domination is strict provided that for some x and K, G2 is strictly better than G1 in respect of K for x.

John:

If one thinks that it is incompatible with divine goodness that he create a world than which there is a better, then one must deny that God could have created w1. In another thread, we've already had a long discussion of the idea that God cannot create less than the best. The idea is incompatible with traditional theistic notions of the gratuitousness of creation, and besides is surely unacceptable if, as seems very plausible to me but not to Mike, there is no best of all possible worlds.

If there is no upper bound on the amount of joy a creature can have, and if it is a necessary truth that the joy of a creature is good and finite at any given time, it follows that any world can always be improved by having more joy given to creatures.

Alex
This is probably way off the mark, but why talk of 'possible worlds' that God could create instead of simply focusing on the one He did create if He exists? It seems to me that the concept of possible worlds creates more problems then it solves. I understand how it relates to ideas of 'necessity,' 'necessary,'and 'contingency,' etc., but the fact still remains; we live in this world as it is with evils and good within it. The question is, would a completely good being create this world. The idea of this being the best of all possible worlds or there not being a best of all possible worlds misses the point that this is the world that we live in and we seek to understand why it is as it is and how we fit into it. Your example of w1, w2, etc. is not about possible worlds but is simply a description of a simlified version of this world as it unfolds. So the question remains, God, if He exists, created this world so how do we explain what appears to be gratuitious evil? Compensation does not appear to work.

John:

I agree that compensation by itself doesn't work. But I also don't know of a good even prima facie case that this world is not one a perfectly good God would make once one factors in the values of significant free will, soul building, simplicity and elegance of laws, virtues that are logically dependent on evils, punishment, overall esthetic considerations, etc., as well as compensation.

If the consequentialist principle is true, then w2 is permissible if w1 is. But w2 includes a gratuitous pinprick. Ergo, etc.

Sorry for the delay. I had no connection available. Maybe I can pick up one thread here. I will get to the rest later, esp. the stuff on incommensurable worlds.

I'm trying to follow the argument above. w1 includes some gratuitious evil. It's unclear to mem but presumably, so does w2. Could God permit w2? No, he couldn't. And that is true under the assumption that some consequentialist principle is the true moral principle. The reason God cannot actualize w2 is not because it is not valuable. The reason is because there is evil in w2 that is unnecessary for the good in w2. To put it another way--consequentialism or not--God cannot actualize a world that contains a single instance of evil that is not necessary to some greater good (or to the prevention of some greater evil).
Now that is just the standard position on permissible evils. What you have said so far gives me no reason to abandon that view. On the other hand, I do not deny that there are reasons to abandon it apart from this discussion.

Dear Mike:

My claim in the argument you're responding to is that the following propositions are not all true:
1. If God can actualize w1, and w2 is better than w1, then God can actualize w2.
2. God can actualize some world.
3. God cannot actualize a world containing a gratuitous evil.

I stand by this claim. Now you can certainly uphold (3)--that claim certainly seems plausible. But if you uphold (3), then you have to either deny (1) or (2), if my argument holds.

Let me repeat the argument. Suppose (1) and (2) hold. Let w1 be a world that God can create (by (2)). Let w2 be just like w1 but where some individual gets both a pinprick and right after it some immense pleasure that he does not get in w1. Then, w2 is better than w1. By (1), God can actualize w2. But w2 has gratuitous evils. Hence, (3) is false. thus, (1), (2) and (3) are not all true.

Now, (1) is just the consequentialist principle I cited. Thus, if (1) holds, only (2) can be denied. But (2) is immensely plausible. Hence, the real choice is whether to deny (1) or (3). I myself deny (1).

Alex,

Couldn't affirming 4 provide a way to affirm 1-3:
4. No world containing a gratuitous evil is better than any world lacking gratuitous evil.

Or:
4'. Any world containing gratuitous evil is worse than any world lacking gratuitous evil.

If one affirms 4, then one has a way to reject any substitution of 1 where w2 is a world with gratuitous evil. Since the problems arise only when a substitution for w2 is a world with gratuitous evil, the inconsistency is alleviated.

So, it seems to me that 1-4 are a consistent set, and that one needn't reject one of 1-3.

Tim:

Yes, but I have an argument against 4, which I gave in an earlier comment. :-) Here's a version of it. One would be better off in getting a very small pinprick followed by a very large extra bunch of pleasure. Everybody and everything else would be at least as well off (assuming there were no further results of the pinprick, say because one was made more annoyable).

We then need the comparison principle that if two worlds w1 and w2 are such that for all x, w2 is at least as good for x as w1, and if there is an x0 such that w2 is better for x0 than w1 is, then w2 is better than w1.

This argument makes two assumptions that can be questioned:
1. The above comparison principle, which perhaps does not do justice to global considerations, such as the elegance of a world without evil.
2. It is better to have a tiny pinprick followed by a great pleasure than to have neither.

But if you uphold (3), then you have to either deny (1) or (2), if my argument holds.

Alex, I am becoming disoriented. I thought we were talking about the claim in (I) that you say is very likely to be true?

(I) Let x be an actual person who suffers the set E of finite evils in his earthly life. Let C be any possible career of x. Then there is a possible career C* such that: (a) C* includes x's suffering the evils from E, (b) for aught that we know, C* is x's actual career, (c) it is better to exist with career C* than not to exist at all, and (d) x would be better off by an infinite amount in having C* instead of C.

Setting that aside, I thought the consequentialist principle stated this (as you say above),

This is tantamount to a denial of the consequentialistic principle that it is always permissible to do better than something that is permissible.

But now I find that the consequentialist principle stated this way,

If God can actualize w1, and w2 is better than w1, then God can actualize w2.

The initial formulation of the consequentialist principle does not have the implication you cite. It is consistent with (2) and (3), as far as I can see. I can hold that there is a world that God can create w1. Let w1 be a world without gratuitous evil. I can agree that there is a better world w2 that God can also create. w2 is better than w1 where w2 also has no gratuitous evil. And I can agree that God can actualize no world with gratuitous evil. So those are consistent.

I'm not even sure where to find this argument in the discussion above. Where is it again?

My claim in the argument you're responding to is that the following propositions are not all true:
1. If God can actualize w1, and w2 is better than w1, then God can actualize w2.
2. God can actualize some world.
3. God cannot actualize a world containing a gratuitous evil.

Alex
But what about W3 - the world that has great pleasure but no pinprick? Certainly a theistically defined God could have created this world. You may be correct that there is no best possible world, but certainly W3 is better then W2 which is better then W1. You need to provide an explanation for W2 being better then W3 if your claim that gratuitious evil is permissible from God's creative standpoint is warranted.

Alex,

I was afraid that you had already given an argument against 4 in the above comments. I read them quickly a couple days ago, and didn't remember it. Sorry about that :)

But, I don't find the argument you give satisfactory.

Suppose A sees that something evil is going to happen to B imminently. A has the ability to stop it, A has no good reason to let it happen, and, caring about B, A would rather B not suffer evil, provided that it isn't necessary for a greater good (and it isn't for a greater good). The evil is gratuitous. If A lets B suffer the evil, then A is doing something immoral.

God is always in the position A is in in this example with respect to gratuitous evils. So, if God allows a gratuitous evil, then God is doing something immoral (isn't that why we think God can't allow gratuitous evils in the first place?)

So, I deny the claim you make that there is a world W containing gratuitous evil where everyone besides the sufferer is at least as well off in W as they are in the closest world without that suffering. In any world with gratuitous suffering, God is immoral, so he is worse off in that world than he would be in a world where he didn't allow gratuitous suffering.

The comparison principle you provide is vacuously true concerning worlds with gratuitous evil, since a conjunct of the antecedent is false in such cases--w2(the world with gratuitous evil) is never as good for x (God) as w1 (the world lacking gratuitous evil). God is always worse off in w2.

Incidentally, that's why 4 is true. No world with gratuitous evil is better than any world without gratuitous evil because worlds with gratuitous evil are worlds where God is immoral. Such worlds are either: 1) impossible or 2) so bad that they are below the minimum threshold of actualizable worlds. In either case, such a world is worse than a world that God can actualize.

Mike:

I was referring back to the argument in my June 21, 2007 3:32 PM comment. You had argued that the consequentialist argument was unacceptable because it did not provide evidence against the claim that there is gratuitous evil. I pointed out that given the consequentialist premise, it follows that God can create a world with gratuitous evil, so the objection in your June 21, 2007 1:26 PM posting did not apply. That said, I suggest we close the discussion of the consequentialist premise, since it doesn't really get us closer to the truth given that (I assume) both you and I reject it.

Tim:

That's a neat move, but I don't know that it does the trick. Consider someone who argues as follows against the consequentialist's idea that one may kill one innocent person to save two others as follows: "Such an action is immoral, because it violates the principle that it is wrong to kill innocent people. Therefore, the consequences of the action are overall negative once you factor in the badness of doing an immoral action." Such a move would beg the question, because the consequentialist is, obviously, rejecting the claim that the said killing is immoral.

In general, it seems to me that a consequentialist principle in order to be non-trivially applicable cannot take into account the moral value or disvalue of the action under evaluation itself. Rather, that action is evaluated in terms of the value or disvalue of everything else. Thus your argument agrees with the letter of (1), but not with the consequentialist spirit. :-)

To assert the principle about the duty to prevent evil seems to be to beg the question the against the opponent who is arguing that a perfect being can tolerate even gratuitous evils.

Nonetheless, what you have is still a splendid way of showing that (1), (2) and (3) can be asserted consistently, though not by someone who is led to (1) by consequentialist considerations.

But the discussion is, again, moot, and I shouldn't have prolonged it, since none of the discussants here have, I think, that much in the way of temptation to consequentialism. And it may be that my intuitions about what motivates people to consequentialism are off, since I am, after all, not a consequentialist.

John:

Ah, but if God created W3, the perfectionist like you could still complain that God didn't create W4, which is just like W3 but contains a pinprick plus a great pleasure. And if God created W4, the perfectionist would complain that God didn't create W5, which is just like W4 but contains the great pleasure but no pinprick. And the game continues.

I think these considerations do show that the fact that a world is not the best possible, or even the best possible for the beings in it, cannot tell against the hypothesis that God exists.

Consider the following putative Principle of Perfection:

(PP) A perfect being will not create a world than which there is a better.

We have good reason to reject (PP), since any being whose character is such as to prohibit them from creating than than which there is a better is a being whose character is imperfect, because such a being's character prevents the being from creating at all. Why? Because for every world there is a better.

So we should reject (PP), and allow that a perfect being can create a world than which there is a better. And now we can ask some neat questions. Consider the collection W of worlds, possible and not, that satisfy all the requirements for creatibility by a perfect being (whatever these requirements might be) except perhaps their compatibility with the goodness of a perfect being (this is kind of a vague definition, and given divine simplicity it may in the end not work, but I want this just an intuition pump). Let G be the subcollection of W consisting of those worlds which are compatible with the goodness of a perfect being.

We can now ask about the properties of the collection G. Plausibly we have the following axioms:

A1. G is non-empty.
A2. If w is in G, while w* is better than w and w* contains no gratuitous evils or (other?) deontological violations by the perfect being, then w* is in G.

I think we should reject:
R1. If w* is better than w, then w is not in G. (This is the PP.)
R2. If w is in G and w* is better than w, then w* is in G. (This is my consequentialist principle.)

I think we should accept the following:
A3. If w in W contains no evils, then w is in G.

One consideration in favor of A3 is the traditional theistic idea that God plus creation is no better than God by himself, and so maximal goodness is always there.

We may have to accept some version of:
N1. If w contains gratuitous evil, then w is not in G.
But there will be issues with the definition of gratuitous evil. One question is whether we shouldn't qualify this as "non-arbitrary"--there is a good discussion of this in Graham Oppy's recent book, where he notes that it may be that a certain good requires an evil, but there is no least evil that will suffice for it. We need some way to finesse that. Another question I will raise in a separate post.

Alex
I am not a perfectionist (although I may at times sound like one). My position is that if God exists He can do what He wants. My problem is that if God creates a world then from His creation what can we ascertain about His nature. The existence of the world is not an a priori truth, it is contingent on the fact that it is what it is and if part of what it is is that it was created, then this suggests that it might not have been created. So from the description of what exists my issue is how we derive any concrete knowledge of what this being is like other then it has the knowledge and ability to create this world.

Our knowledge of God cannot be derived from some definiton because the question then would be does what is being defined exist. The ontological argument fails not because it is not soound (it may very well be so) but because even if we define God as being perfect and mainntain that the existence of such a being would be more perfect then if it did not exist, it still remains a possiblity that a perfect being does not exist. What we do know is that from the very fact of creation, if it is a fact, is that the being that created it had the knowledge and ability to create it. It does not tell us what the moral nature of that being is, or even if that being still exist. Maybe the act of creating the world killed the being creating it (like some mothers in childbirth).

So the problem of gratuitous evil is not that God does not exist, but that we cannot derive any knowledge of that beings moral nature if He created the world and it contains this type of evil. If it is necessary for a greater good then it is not gratuitious. But that still does not help us determine the moral nature of the being who created this world, if it was created.

Dear Alex,

Actually, I was wondering what earlier argument the June 21, 2007 3:32 PM comment alluded to. I guess I didn't see its importance to the main argument you were defending. But you're right, I do deny the consequentualist principle as you describe it in (1) of that post which amounts to the claim that (where I quantiy over worlds, 'A' is actualizes, 'M' is possibility,

1. (Vx)(Vy)(M(GodAx)->((y > x)->M(GodAy))

(1) claims that God can actualize any world better than a world he could actualize.

I do not reject the consequentialist principle as you describe it in the first paragraph of June 18, 2007 11:01 AM. The version here states that,

2. (Vx)(Ey)(M(GodAx)-> ((y > x) & M(GodAy)

(2) claims that God can actualize some world better than a world he could actualize. I guess that sums it up.