# God and Evil

Back in the comments section of my post on Giving Thanks, an old college friend and I are discussing the age-old problem of why God permits suffering and other evils.  This is a serious problem; in my view the Argument from Evil is the only really formidable positive argument for Atheism.  (By a positive argument for Atheism, I mean something that provides specific evidence against God's existence, rather than merely making the negative claim that there isn't enough evidence for Theism to believe it.  In order to show that Christianity is plausible, both claims must be addressed.)

The conundrum is famous: If God is the All-Knowing, then he knows what things are evil, if he is the All-Powerful, he should be able to prevent them, and if he is the All-Loving, then he will want to prevent evil.  So why is there evil?

The only way to solve the problem is to postulate the existence of some good thing which cannot exist unless evil either exists, or is at least possible.  (Common "defences" might refer to putative goods such as free will, the opportunity for humans to exercise virtues, the orderliness of the universe, an afterlife of a sort that depends on people having had certain experiences, etc.)  If the good is such that it is logically impossible to get it without (possibly) getting the evil too, then the defence would be successful, since when we say that God can do anything, we don't mean that he can or would create a logical contradiction.  (As C.S. Lewis says in The Problem of Pain, "Nonsense remains nonsense even when we talk it about God.")  I'm not going to attempt a detailed defence here, but I do want to make some general points about the Argument from Evil.

My first point is that God's omniscience actually makes the Argument from Evil weaker, not stronger.  The reason is that we humans are not omniscient.  If we are ignorant, there's no particular reason to assume that we know what is the morally best way to run a world.  Suppose that you wrote down a list of all the things you regard as good (happiness, knowledge, beauty, whatever).  Suppose you figured out a way to weight all of these factors numerically—of course, there's no way we could ever agree on how to do this, and I'm not convinced it even makes sense, but let's run with it—so that you could assert that some possible kind of universe (call it $U$) is optimum: the best possible.

[Note for experts: my kinds of universes $U$ here aren't exactly the same as the "possible worlds" discussed by analytic philosophers.  If the best possible kind of universe contains something like free will or nondeterminism, there will be multiple "possible worlds" $W_1, W_2 \ldots$ consistent with the same overall plan $U$ of the universe, some of which may be morally better or worse compared to the others.]

Now if God knows about even a single kind of goodness that we are ignorant of, or if he weights the various kinds of goodness differently than us in any way, then of course God will view some other kind of universe $U^\prime$ as best.  It seems infinitely unlikely that $U = U^\prime$ just by coincidence, so it seems to be almost certain that the universe will appear to us to contain evils that we can't explain.  One can argue about whether this is a sufficient explanation, but it's definitely something that has to be taken into account.  The idea that a superhuman entity which created the universe will see things exactly the way we do is absurd:

“For my thoughts are not your thoughts,
neither are your ways my ways,”
declares the Lord.
“As the heavens are higher than the earth,
so are my ways higher than your ways
and my thoughts than your thoughts.”
(Isaiah 55:8-9)

The second point I'd like to make, is that the Argument from Evil has emotional force as well as intellectual force.  Atheists tend to get annoyed when Theists suggest that Atheists don't believe in God because they resent him.  I've certainly seen plausible cases of this, but I don't want to speculate that all Atheists are this way, since I don't like making unfounded accusations about individual people's characters.  (Maybe that's why my Politics category only has one post in it so far.)

Nevertheless, leaving the Atheists aside for a moment, I think I can say from an examination of my own heart, and conversations with other people, that it's easy to carry an unconscious grudge against God for various real or imagined grievances in our lives, or the lives of those we care about.  Even if we have no grudge, there can be a deep sense of pain from all the kinds of grief that we don't understand.

So the Argument from Evil carries emotional force as well as intellectual force.  There's no necessary reason why an intellectually satisfying answer should be an emotionally satisfying answer, or vice versa.  One should bear this in mind when evaluating the intellectual arguments, since we may be asking from an argument something that no argument can do.

Finally, I believe that Christianity has resources for addressing the Argument from Evil which don't exist in generic-brand Theism, or indeed in any other religion.  It's much too simple to say that the existence of evil contradicts Christianity, when in fact the most basic doctrine of Christianity logically implies the existence of evil.

The basic doctrine is that 1) we human beings are wicked and deserve punishment, and that 2) in order to forgive us, God became an innocent human being and allowed himself to be tortured to death by us, and that 3) this act provides us with spiritual healing now, as well as physical immortality for all eternity.  Now regardless of whether you like this idea, even if you find it implausible or downright incomprehensible, you must admit that it's an idea about how God relates to evil, and uses it for the sake of good.  If there were no such thing as innocent suffering, Christianity wouldn't even be possible.  If Christianity is true, then God has arranged things so that the most important thing that ever happened was a horrible but redemptive evil.  All other evils, we view in the light of the Cross.

I am a postdoctoral researcher studying quantum gravity and black hole thermodynamics at the Institute for Advanced Study in Princeton. Before that, I read Great Books at St. John's College (Santa Fe), got my physics Ph.D. from U Maryland, and did my first postdoc at UC Santa Barbara.
This entry was posted in Theology. Bookmark the permalink.

### 27 Responses to God and Evil

1. almkglor says:

The simplest explanation for evil is: the universe (or God, whatever), does not care about evil at all, but simply allows both good and evil to occur. So there's just one doctrine: (1) the universe does not care.

Any explanation more complicated than that has to give more accurate predictions than that. Christ's death is not a sufficient prediction, because it's a past event, and the people who made that prediction (i.e. the Israelites) were the same people who made it happen (i.e. the Israelites).

2. almkglor says:

It seems infinitely unlikely that U = U' just by coincidence, so it seems to be almost certain that the universe will appear to us to contain evils that we can't explain.

I think we should be more precise here.

Call the "humanly ideal" universe Uh. Call the "real" universe Ur. Observation shows that Uh != Ur, i.e. the problem of evil.

Theodicies try to say that, actually, there is a "Godly ideal" universe Ug, and that actually, Ug = Ur (an assertion that must be borne by some argument or evidence).

You point out that due to omniscience, it is very very likely that Uh != Ug.

Now, given Uh != Ur and Uh != Ug, please tell me why I should accept that Ug = Ur.

(I have further arguments regarding probability, since plain Arestotelian logic is insufficient - we need Null-A here).

3. almkglor says:

(continuing previous comment about Uh, Ug, and Ur)

Now, consider a universe composed of one bit. In that case, knowing Uh != Ur and Uh != Ug, the probability of Ur = Ug is 1.0 (consider the case that Ur is 0. Then Uh can only be 1. Knowing that Uh != Ug, we know that Ug can only be 0, which is the same value as Ur)

For a universe composed of two bits, knowing Uh != Ur and Uh != Ug, gives us the probability that Ur = Ug is 0.3333 (or 1/3) (consider he case that Ur is 00. Then Uh has 1/3 probability each of 01, 10, 11. Assuming Uh is 01, then Ug itself has 1/3 probability each of 00, 10, 11. So in that limited case, Ug = Uh 1/9 of the time (probability of Uh = 01 times probability of Ug = 00. But there are three possible cases of Uh, so 3 times 1/9 yields 1/3).

For a universe of N bits, 1.0 probability Uh != Ur (real observation) and O probability Uh != Ug (omniscient unlikelihood), then the probability (assuming O is 1.0) of Ur = Ug is 1 / ((2 ^ N) - 1). If O is non-1.0, then that is multiplied to that base probability.

We know for a fact that N (the size of the universe) is very large. According to this page, the number of bits in the universe N is 10^90.

The base probability of Ur = Ug without the omniscient unlikelihood is 1 / (2 ^ N). Adding omniscient unlikelihood does not seem to improve the base probability by much.

(I have more to contribute, this time regarding the probability of omniscient unlikelihood O)

4. Aron Wall says:

almkglor,
You are of course perfectly correct that the nonexistence of a benevolent God is the simplest explanation for the existence of evil. The belief that there is a beneovolent God who permits evil for some good purpose is logically possible, but more complicated. I agree. That, of course, is why the Argument from Evil is typically thought of as an argument for atheism.

The remaining questions are: 1) Just how powerful of an argument for atheism is it? and 2) are there any positive arguments for theism which outweigh the Argument from Evil, and therefore make theism more plausible than atheism?

The purpose of this post was to address (1) only. Think of it as damage control, if you like. I'm trying to explain why the argument from Evil, although it has some force, does not completely settle the issue. In other words, that it is not grossly implausible that $Ug = Ur$, simply because $Uh \neq Ur$. That's the point of my reflections on omniscience and the role of evil in Christian theology. Then, if I can say that there exists evidence for Christianity, which outweighs the evidence against it coming from Argument from Evil, then we should still believe in Christianity. And since Christianity includes belief in a benevolent deity, this in turn would imply that $Ug = Ur$.

If you need me to write this out as a formal Bayesian probability calculation, I can do so. But there's nothing much too it: it's just the ordinary workaday observation that sometimes the evidence for a proposition can outweigh the evidence against it.

Note that (2), what I consider the positive evidence for Christianity, is not contained in the post you're responding to. I haven't really analyzed (2) much yet in this blog, since I'm still laying the groundwork with my analysis of what Science is and why it works. But it will come in due time. You can see my previous reply to you if you want to know what kind of evidence I regard as most convincing.

UPDATE: The number of bits in the universe is irrelevant to the probability calculation. It's related to how any particular license plate you see on a car is very unlikely if the plate was selected randomly, but nevertheless that doesn't disconfirm the "random selection" hypothesis much relative to the "vanity plate" hypothesis, since the odds are small either way.

5. almkglor says:

The purpose of this post was to address (1) only. Think of it as damage control, if you like. I'm trying to explain why the argument from Evil, although it has some force, does not completely settle the issue. In other words, that it is not grossly implausible that Ug = Ur, simply because Uh != Ur. That's the point of my reflections on omniscience and the role of evil in Christian theology.

But the basic analysis above shows that it is still grossly implausible. In short, you've just shown that the probability is 1 / ((2 ^ N) - 1) rather than 1 / (2 ^ N). The difference is 1 / (((2 ^ N) - 1) * (2 ^ N)).

Then, if I can say that there exists evidence for Christianity, which outweighs the evidence against it coming from Argument from Evil, then we should still believe in Christianity. And since Christianity includes belief in a benevolent deity, this in turn would imply that Ug = Ur.

Okay, lemme see if I get this straight: what you're saying is that "I don't know if Ug = Ur, but if we find evidence X, Y, Z, then the probability of Ug = Ur should improve over the prior base rate. In no way can X, Y, or Z depend on Ug = Ur because that would be circular logic." Is that right?

UPDATE: The number of bits in the universe is irrelevant to the probability calculation. It's related to how any particular license plate you see on a car is very unlikely if the plate was selected randomly, but nevertheless that doesn't disconfirm the "random selection" hypothesis much relative to the "vanity plate" hypothesis, since the odds are small either way.

WHAT? Please explain this more fully. In fact, please give me your Bayesian priors and Bayesian analysis.

If what you're arguing is that a Universe composed of a featureless repeating substance is inherently boring and obviously not our world, then please consider the size of the Universe in compressed form, which is an approximation of the Solomonoff induction of the Universe's program. I doubt that that number is still significantly large enough that raising two to that number and getting the reciprocal would yield any probability large enough to be more salient than the probability I will be hit by a train tomorrow.

6. Larry Wall says:

I am amused when people try to infer the quality of the universe
from its quantity. Surely this is a category error. Works of
art come in all sizes, and their quality is largely orthogonal to
how big they are.

As for the quality of the universe, too many people seem to prefer to
maintain a double standard here. If something good happens to us, that's
because we're lucky enough to live in this universe. If something
bad happens, that is somehow God's fault, more so than the good stuff is.

Sometimes I ask such people this question: "If you had a Big Red Button
that would annihilate the universe right now, would you push it?"
Generally (unless they're feeling suicidal) they'll say something like,
"No, the universe is an interesting and beautiful place, why would I
destroy it?" People tend to like the fact that they exist. They think
the universe is on the whole a pretty good thing—but only if it
happened randomly. If a Cosmic Artist invented that same cosmos,
he must somehow be considered naughty.

Most human artists understand that they must serve their work of art,
and that any attempt to impose an external program on the work of
art will reduce its artistic integrity. I believe that the universe
demonstrates just such artistic integrity: we would find ourselves
in a much less interesting story if God arbitrarily "fixed" things
for us whenever we think he ought to. We would be deprived of the
joy of our own character development. I am glad to have been written
into Our Story, even when things seem to be a bit suboptimal from
my perspective.

7. Aron Wall says:

almkgor,
Here's a simple toy model of probabilities to demonstrate why the number of bits in the universe doesn't matter. Suppose that there are $X$ possible configurations of the world obeying certain laws of nature. And suppose---as a simplifying assumption which I don't actually believe, for purposes of illustrating the idea by taking an extreme case---that we have no idea which of these possible configurations God considers to be the best. In that case, if God exists we would consider each of the $X$ configurations to be equiprobable. Now if we observe some specific configuration $x_1$ on this hypothesis $G$ is
$P(x_1 | G) = 1/X$.
You might be tempted to say, that's grossly improbable because X is so large. But now suppose we have some other naturalistic hypothesis $N$ which also does not tell us specifically which configuration of the universe to expect. In this case, the likelihood ratio is the same:
$P(x_1 | N) = 1/X$.
So when we compare the relative merits of $N$ and $G$, we end up with the same likelihood ratio. Hence the probability ratio of $N$ and $G$ is just the ratio of prior probabilities, whatever they are. All else being equal, presumably $N$ wins since it's simpler.

Now obviously, in the real world $N$ and $G$ don't make exactly the same predictions, so this model is too simple. For one thing, it's unreasonable to assume our knowledge of what God would consider "good" is completely inaccurate. Hence the Argument from Evil, which raises the posterior probability of $N$. Conversely, if God does specific things like miracles which are unexpected naturalistically, then this can increase the probability of $G$ relative to $N$. But I don't think the force of either of these arguments will be particularly sensitive to $X$, precisely because the Naturalist and the Theist (if they accept similar scientific theories) will both have the same predictions for most (but not all) of the bits of data in the universe.

8. Dennis Towne says:

You wrote: "But now suppose we have some other naturalistic hypothesis N which also does not tell us specifically which configuration of the universe to expect."

I believe that this is one of the fundamental flaws in your estimates and calculations. Ug specifies one particular configuration, and the probability as listed above is 1/X; god is perfect/omniscient, and has one ideal configuration. However, Un does not specify one configuration. While there may be a perfect 'god universe' configuration out there, the naturalistic explanation has many, many, many possible configurations.

Comparing N and G in this context gives N a huge prior compared to G, and G has a correspondingly huge burden of evidence needed to catch up.

9. Aron Wall says:

Dennis,
In my toy model, it's important that we don't know which of the $X$ possible configurations is $Ug$. If we knew which world was $Ug$, and we observed something different from that, then the probability of $G$ would be 0, not $1/X$. But since we don't know which world is best, both naturalism and theism have exactly the same posterior predictions (in the toy model, not reality).

Or is your point that $G$ has a smaller prior probability because it makes more specific predictions (even though we are not in a position to know what those predictions are)? If so, that's not how Bayesian probability theory works. You don't punish a theory for making specific predictions. Making specific predictions is good, if you're in a position to observe those specific predictions. If you aren't in a position to observe the prediction, then it's a wash, and you can treat it exactly like it didn't make the prediction.

Let me be more specific. Let's suppose we have a deck of 52 cards, and suppose I flip up the card from the top of the deck and look at it. It's the $3\spadesuit$. Let $N$ be the hypothesis that the deck was shuffled randomly before I flipped the card. Let $G$ be the hypothesis that a cat burgler broke into my home when I was asleep, and put his favorite card on top. Clearly $G$ is less likely a priori, but for reasons that have nothing to do with the number of cards in the deck! I wouldn't say that the cat burgler hypothesis was less likely if there were more cards $X$ in the deck.

Now what about the likelihood ratios? If I think I know what the cat burgler's favorite card would be, I've got some data to extract a prediction from. But if I have no clue about his favorite card, then I just go with my prior probabilities, which don't depend on $X$. Agreed?

10. almkglor says:

I am amused when people try to infer the quality of the universe
from its quantity. Surely this is a category error. Works of
art come in all sizes, and their quality is largely orthogonal to
how big they are.

I wonder how much quality you'd find in a blank canvas. Hence the reduction to bits in compressed form.

11. lavalamp says:

The argument only works if you have absolutely no idea what Ug is. But I think for any reasonably well defined god, Ug will be at least constrained. If you really have no idea what Ug is, it's not clear to me that your god is meaningfully defined. The argument that since god is omniscient, we can't expect to predict Ug at all sounds, to me, like an ad hoc construction made specifically to evade this criticism.

A second point, not metioned above: I think that any Ug much different from Uh implies the god is not benevolent or not interested in us. Not that we should ignore the possibility of non-benevolent gods, but that would rule out the christian god, at least as commonly conceived.

Together, these two points make it very difficult for me to imagine a Ug consistent enough with both Uh and Ur.

12. Aron Wall says:

Dear lavalamp,

Yes, my statement that we can't predict Ug at all is a totally ad hoc assumption, and what's more, I don't believe it! That's why I called my probability analysis above a "toy model", i.e. a oversimplified model which one designs to explore a specific issue, rather than to fit the real world. It was designed to illustrate the specific point that the total number of bits in the universe doesn't affect the Bayesian analysis. That's why I went on to say "Now obviously, in the real world N and G don't make exactly the same predictions, so this model is too simple...." and then I described some of the specific ways in which it is too simple.

Ug is certainly quite different from Uh, but that's doesn't mean that God isn't benevolent. Uh is what we think is best for us. Ug could differ for one of two reasons: either (i) because he's not as interested in our benefit, or (ii) because he thinks he knows better what will benefit us. In the second case, Ug could differ quite substatially from Uh without implying that God isn't trying to benefit us. A parent may do things for a young child's benefit that are substantially different from what the child thinks are best, but that doesn't mean they aren't benevolent.

I don't know what the "christian god, as commonly conceived" would do, but if you take the Christian God as conceived by the Bible, then he very clearly states that suffering is one of the things that's good for us. He didn't leave us to speculate on this point. Although Jesus did heal many sick people as a token of his good will towards humanity, he also specifically says that "if anyone would come after me, he must deny himself and take up his cross daily and follow me. For whoever wants to save his life will lose it, but whoever loses his life for me will save it. What good is it for a man to gain the whole world, and yet lose or forfeit his very self?" (Luke 9:23-25. In other words, Christians must be willing to experience something like death by torture, in order to become children of God as Jesus was. (see Heb. 12:5-13)

Perhaps to make things easier for us, God has arranged the world so that great suffering (though in different degrees) is likely to come at some point to all of us, religious or not, whether we seek it or not. So I think that if we're talking about the God of the Bible, then we're in for some pretty heady departures from Uh, or at least from what Uh would be if uninformed by divine revelation.

13. lavalamp says:

Two things in response.

A) It seems like your toy model differs from your real model on exactly the point you constructed it to defend yourself from? In the toy model, you can't tell the difference between Ug and Ur, but in your real model, you can. It seems like the number of (compressed) bits in the universe matters in your real model, but not your toy model. (In bigger universes, you have more evidence.) In other words, your toy model seems like a red herring?

B) Are you claiming that god values suffering as a terminal, or instrumental goal? If terminal-- well, it seems clear to me that makes him/her/it malicious, but that seems to be a pretty non-standard view for christianity*, so I doubt you hold it. If instrumental, then for that to be acceptable, it must be in service of a terminal goal that is a greater good than the suffering, and cannot be reached in any other way. I think there's a very low probability for such a goal existing.

*Caveat: I'm from a protestantish background-- I'm guessing you're more Anglican/Catholicish, and I know less about those traditions, so I could be wrong here. But "desires human suffering as a terminal goal" sounds so clearly malicious that I have a hard time believing it's standard anywhere...

14. g says:

Larry -- the proposition "the world is, on the whole, a really bad thing, worse than outright nonexistence" is quite different from the proposition "the world is much worse than should be expected if it's the creation of a supremely good and supremely powerful being". The second of those is what's claimed by people making an Argument From Evil; the first is the one you're giving evidence that atheists don't agree with.

Aron (on the possible not-badness of suffering) -- 1. in so far as Christianity (or a particular variant thereof) says that suffering is a Good Thing, it thereby becomes less plausible. 2. It seems to me that the commonest Christian view of this sort is that suffering is (sometimes) good by being *redemptive*; that is, sometimes something is very badly wrong, more badly wrong than any amount of suffering, and suffering is necessary in order to put it right. In so far as suffering is necessary (or outright good) for that reason, what we've got is a universe with something in it that's even worse than skeptics say all that suffering is. This doesn't seem like an improvement from the point of view of someone arguing for the existence of a good and powerful god. 3. Suffering is by no means the only bad thing in the world, and it is not the only bad thing to which a well-crafted Argument From Evil will appeal. For example, if traditional forms of Christianity are right then it is a very, very bad thing for a person to be badly wrong about God's existence, character and purposes, and many, many, many people are in fact badly wrong about those things.

Aron (on Ug, Ur, and Uh) -- I quite agree when you say "it is not grossly implausible that Ug=Ur, simply because Uh≠Ur", but I think that's responding to an argument different from the one skeptics are actually making. It's not the mere fact that Uh≠Ur that's evidence against the existence of a good and powerful god who cares about our affairs, it's the extent and nature of the differences between Uh and Ur. I think one can make a very good argument that those do make it grossly implausible that Ug=Ur, though of course they don't amount to a watertight proof that Ug≠Ur.

15. Aron Wall says:

lavalamp,
A) Even in a realistic model, the total total number of bits in the universe is only relevant if we have strong opinions about how those bits ought to be, and observe them not to be that way, conditional upon what we've learned from previous bits. So the number of bits in e.g. the Andromeda Galaxy is not relevant, because there's no strong argument that God, if he were really benevolent, would have created it with different properties than we observe. (Things might be different if we observed alien life there, but we haven't yet.)

What really matters is the a priori implausibility of the assumption "God has some good reason for permitting lots of suffering/evils". I agree that this is somewhat implausible, but it's important to realize that we only takes this hit once. Suppose we take the Holocaust as an example of grevious evil. You may find it difficult to believe that God would have a good reason for permitting it. But suppose he does. Conditional on that, the genocide in Rwanda doesn't provide much additional evidence, because we already know from the Holocaust that God has some general sort of reason for permitting genocidal evils. This conditional-probability analysis applies even if we don't know what that reason is.

B) I'm only claiming that God values suffering in an instrumental way. As the author of Hebrews says, "Let us fix our eyes on Jesus, the author and perfecter of our faith, who for the joy set before him endured the cross, scorning its shame, and sat down at the right hand of the throne of God."

Is it improbable that such a terminal joyful goal should exist? I don't think it's all that improbable---even from the perspective of purely human wisdom uninformed by divine revelation---that suffering provides opportunities to build character, and that character is one of the most valuable things to have in life. The probability of this being a viable purpose for suffering is increased greatly, if this world is a preparation for the next life, rather than being the end goal all by itself. And of course that's what Christianity says it is.

Taking these things into consideration, I think the Argument from Evil decreases the probability of theism by maybe two to four orders of magnitude. That's a blow, for sure, but not an insurmountable one if there's compensating positive evidence. And part of the point of my top post was to argue that Christianity does better than generic theism, because $P\mathrm{(evil\,|\,Christianity)} > P\mathrm{(evil\,|\,theism)}$.

C) I'm actually of Protestant origin too--specifically the Wesleyan/Holiness tradition. That was originally an offshoot of Anglicanism, so maybe that's what you're detecting, although I'm pretty ecumenical in my reading. I grew up in the Church of the Nazarene, and I'm currently attending the Free Methodist Church of Santa Barbara.

16. lavalamp says:

A) A properly compressed bitstring representing the universe is probably going to look a lot like a program expressing a set of physical laws, plus constants/initial conditions, etc. It probably won't even have a separate term for the andromeda galaxy. I agree that, given the rest of the universe, the andromeda galaxy gives us very little if any new information. I agree that the total number of bits isn't a good measurement unless you compress it. (This assumes the universe is computable, which seems safe at the moment.)

B) OK, I personally would put the range more like 3-8 orders of magnitude, but I can probably live with 2-4. However, I don't agree that christianity does better than generic theism, because generic theism isn't committed to a benevolent god. It seems to me that basically all religions that posit a benevolent deity and want to try to seem reasonable are going to have to end up using the justifications christianity uses (free will defense, etc), so the fact that christianity has this reasoning doesn't seem like much evidence one way or the other to me.

C) Hm, my spidey sense is rusty. I think the few occurrences of "St." threw me off. The last church I went to before it all stopped making any sense was actually a (non-typical) Wesleyan.

17. Aron Wall says:

g,
I actually agree with a lot of what you say, but I have a few small points in reply:

1. in so far as Christianity (or a particular variant thereof) says that suffering is a Good Thing, it thereby becomes less plausible.

Yes, to some extent. But I'm trying to argue that Christianity does more than just assert that suffering is a good thing, it also shows us through specific examples, and connections to other ideas, a way in which this could be the case. So I find the Christian view of why suffering exists to be much more satisfying and plausible than the generic theism view. It still contains some mysterious elements. But because of the Incarnation these mysterious elements are not totally hidden in God's mind, instead they have come into the world and have become visible.

2. It seems to me that the commonest Christian view of this sort is that suffering is (sometimes) good by being *redemptive*; that is, sometimes something is very badly wrong, more badly wrong than any amount of suffering, and suffering is necessary in order to put it right. In so far as suffering is necessary (or outright good) for that reason, what we've got is a universe with something in it that's even worse than skeptics say all that suffering is.

I'm not sure this quite follows, since the net result of the thing going wrong, and then being fixed, could conceivably be better than the thing never going wrong in the first place. (A truly effective fix would have this property.)

And the suffering may be partly necessary for developmental reasons (something that turns a child into an adult) in addition to being redemptive (something that turns a bad person into a good person). That's why the author of Hebrews says that Jesus was "made perfect [i.e. complete] through suffering" even though he didn't sin. (See Heb. 2:10-18)

Also, although redemption through suffering does imply the existence of something worse than the suffering itself, there may be resources to explain the existence of that greater evil which wouln't apply to the lesser evil. For example, a free will defense is a lot more plausible as an explanation for moral corruption, than as an explanation for why people get cancer.

3. Suffering is by no means the only bad thing in the world, and it is not the only bad thing to which a well-crafted Argument From Evil will appeal. For example, if traditional forms of Christianity are right then it is a very, very bad thing for a person to be badly wrong about God's existence, character and purposes, and many, many, many people are in fact badly wrong about those things.

While this does count as a bad thing in the here-and-now, it's not necessarily a complete loss so long as the person eventually (either in this life or the next) discovers the truth and comes into loving communion with God. Some people might even gain valuable experience from making decisions on their own for several years, before they realize that God has a claim on their life.

But I agree with your main point here. If Christianity is true, then Hell is a much worse thing than any amount of earthly suffering. Although this could be used to gain some traction vis-a-vis the problem of suffering (obviously God should be more ruthless in causing suffering if it is needed to prevent damnation), it leaves a much worse evil to be explained. And this is something which is explicitly stated to be against the Christian God's goals, since "God wants all men to be saved and to come to a knowledge of the truth" (1 Tim 2:1-4). However, since God is just this can only happen to people who deserve Hell and choose it by their actions. And the possibility of a bad outcome follows logically from the claim that true joy requires ethical character. If God declines to force ethics on an unwilling subject, then his two remaining choices are annihilation (i.e. capital punishment) or some version of Hell (i.e. life imprisonment). It's not obvious to me which of the two remaining choices is better, but apparently God has picked the latter.

18. Aron Wall says:

lavalamp,
B) I think 8 orders of magnitude is a little high, but we're obviously in roughly the same ballpark as each other since you accept 3 as the lower range. And I guess I actually meant theism-with-a-benevolent-God-but-not-a-specific-organized-religion.

I've noticed however that very few monotheists believe in an active but nonbenevolent deity. At least, most people who accept the Argument from Evil end up disbelieving in God altogether, rather than accepting a nonbenevolent God. Is that because the thought of a nonbenevolent God makes people too uncomfortable, or is it that most atheists implicitly accept some argument that a God must be good, if he exists? (Polytheists often believe in malevolent deities, but their gods are conceived of as super-powerful but contingent beings, not the unique Creator of Everything)

The thing that's in Christianity but not other monotheistic religions is the Crucifixion as a central part of the theology of how God relates to the world. This considerably changes the emphasis of the standard defences. (As an aside, I've noticed that other religious traditions tend to teach that all suffering is deserved, while Christianity emphasizes the possibility of innocent suffering.)

C) The "saint" prefix is not a typical Wesleyan practice. It comes from this blog's highly eccentric canonization policy.

19. g says:

Aron: on 1, I understand but I think we have very different ideas about how "satisfying and plausible" Christianity's reasons for suffering are. For instance: yes, it gives specific examples, such as the alleged connection between the suffering of Jesus on the cross and the redemption of (perhaps some favoured subset of) the world, but it doesn't give credible explanations of how suffering is actually beneficial in these instances. (There are many mutually incompatible Christian "theories of the atonement", and none of them really makes much sense to anyone other than Christians.)

I also want to make a general point about the idea that Christianity has extra resources for dealing with evil that theism-as-such doesn't have. Clearly, Pr(Christianity|evil) can't be higher than Pr(theism|evil), though it can certainly *feel* otherwise (see: conjunction fallacy). Of course considering Christianity can draw attention to some justifications for evil (note: here "evil" is short for "the generally agreed facts about the existence, quantity, distribution, etc., of various kinds of bad things in the world) that might otherwise have gone unnoticed; but please don't neglect the possibility that what's actually going on is that you're pushing improbability from one place to another. That is: Christianity is *much, much more specific* than theism-as-such and its prior probability is correspondingly much lower. And every amazing thing you ascribe to it -- it explains evil! it justifies belief in objectively real moral values! it provides a non-question-begging explanation for the orderedness of the universe! etc. -- correspondingly reduces its prior probability when you build it into your definition of "Christianity" (or of "God" or whatever). The only way to get a real gain from this sort of manoeuvre is if you assume something that really isn't all that improbable, and get from it conclusions that otherwise would have been really improbable. That *might* work out in the case of Christianity, but it's a case you actually need to make and it's far too common to see it just swept under the carpet.

On 2, of course a greater final good might justify all that bad stuff. (Isn't that idea taken for granted when we have this sort of discussion at all?) My point is just that the worse the bad stuff is, the greater the justifying final good has to be -- and if the suffering is good because it's the way around something even worse, well then that something-even-worse must in fact be even worse, and the final good must be good enough to justify it.

On 3, for now I'll just note that "most people deserve hell" is itself a claim that's so implausible that any version of Christianity that makes it is thereby made much less plausible than it would have been without that claim. Oh, one other thing: you equate annihilation with capital punishment and hell with life imprisonment, but hell is in fact *infinitely worse* than life imprisonment whereas annihilation is *pretty much exactly the same thing as* capital punishment. Which makes it pretty weird, to my mind, to suggest (as I think you're doing) that the two look about equally bad and that it's not terribly surprising if God chooses "life imprisonment" rather than "capital punishment".

20. Aron Wall says:

Regarding whether the Atonement makes any sense at all, that subject will justify an entire post all by itself. I think it may boil down to whether or not one is willing to accept a partially comprehensible explanation for what's going on. It sure helps if one already has basic concepts such as "sacrifice", "holiness", "atonement", and so on in ones worldview. Christianity started in a time and place where most people, even pagans, had these concepts. It must now engage in dialogue with people who don't have these concepts at all, except maybe when they read fantasy novels.

When I say that Christianity has extra resources for dealing with evil, what I mean is that P(evil | Christianity) > P(evil | benevolent theism). Obviously, the probability of Christianity cannot rise above the probability of theism. It can, however, be the case that Christianity is the most plausible form of theism given some set of evidence.

Christianity is much, much more specific than theism-as-such and its prior probability is correspondingly much lower.

I agree. This makes it quite hopeless, I think, to postulate Christianity speculatively simply to explain away the Argument from Evil. In order to get to Christianity specifically we need evidence specifically for some of the special features of Christianity, such as the historcal argument for the Resurrection.

I don't endorse the claim that "most people deserve Hell". At least, not unless we use a very restricted and circumscribed definition for the tricky word "deserve". I do think that (i) without divine help none of us can reach the moral standard which God wants us to attain, and (ii) all of us (except young children) have had moments in which we chose a standard of behavior which, if we continued along that path without repenting, would eventually corrupt our character irretrievably, and in that sense we are all in danger of Hell. But (i) and (ii) by no means imply that we deserve Hell, because there is an enormous mitigating factor: we can't really help being that way. As a result, people can only strictly deserve Hell when they continually reject and resist God's way of salvation, for so long that they get stuck in that position forever. And by "reject and resist", I mean something a lot stronger than "honestly don't believe there's evidence for".

I do think there's a small minority of people here on Earth who are pretty much damned already, although it's not my place to judge and we have to hold out hope for even the worst of sinners. There's some terrifying case studies in St. M. Scott Peck's The People of the Lie. Another book which is helpful for seeing how ordinary people might choose Hell is St. Lewis' The Great Divorce.

When you say that Hell is infinitely worse than life imprisonment, I think you're begging the underlying question regarding whether existence in Hell is better than non-existence. If existence is better than non-existence even for those in Hell, then it seems to me that annihilation isn't better, no matter how long Hell lasts (if it exists in time at all).

Note: although you're free to go on using *'s if you like, the comments box does support italics and certain other html tags.

21. g says:

Ignoring the fluff and getting straight to the central point: yeah, I know I can use italics instead of *asterisks*, and in a couple of places I have. It's just that (1) sometimes I'm just too lazy to type the extra characters, (2) I actually quite like the look of the *starred* style, and (3) I think maybe the two forms of emphasis produce slightly different effects, though I'd have trouble putting my finger on exactly what the difference is, and sometimes I use both for that reason.

OK, back to trivia like whether God exists and whether anything can justify hell.

I agree that a proper defence of the idea of the Atonement -- or for that matter a proper attack on it -- would take a post in itself. (As it happens, my own issues with the idea don't stem from not having the relevant concepts in my brain; I was a Christian for thirtysomething years, and not a particularly oblivious or untheological one either, but I'm sure you're right that plenty of people are in that situation and that anyone wanting to make Christianity genuinely credible to such people has some work to do.)

I understand that your claim is Pr(evil|C) > Pr(evil|T); I just wanted to point out (1) that the thing that actually matters is Pr(C|what-we-know) rather than Pr(anything|C) as such, and (2) that it's easy to concoct theories with Pr(known-observations|theory) very close to 1, simply by building lots of those known observations into the theory, and that merely doing that doesn't actually give the theory any extra credibility overall because it just pushes improbability into the prior probability of the theory, and it's easy to overestimate how much you're raising Pr(observations|theory) and underestimate how much you're lowering Pr(theory). I'm sure you understand all this already, so I'll shut up about it now.

I take your point about deserving hell, so let me refocus a little. Traditional versions of Christianity imply something like the following claim: The majority (perhaps the overwhelming majority) of the human race, if God behaves optimally towards them, will most likely end up in eternal torment and deserve it. I don't know what your moral intuitions say about that, but mine are pretty unenthusiastic.

I'm not begging the question of whether existence in eternal torment is worse than nonexistence. I'm appealing to the obviously-correct answer to that question (namely "ye gods, of course it is") in order to address the quite different question of how credible the traditional doctrines of Christianity are.

I have read "The great divorce". It may very well be that otherwise-fairly-decent people can be put into situations in which they will "choose hell". If so, then it seems to me that a being of vast power and benevolence might reasonably be expected to keep them out of such situations as far as possible. And that the more carefully one thinks about CSL's portrayal, the less credible it looks -- at least, if one takes it as a serious proposal of a possible state of affairs in which "eternal damnation for a large percentage of the human race" is a reasonable outcome. (It's not perfectly clear to what extent it is intended as one, and for the avoidance of doubt I am making no claims about CSL's precise intentions.)

For instance, the denizens of CSL's kinda-sorta-pre-hell seem to be almost entirely oblivious of where they are and what their likely fate is. Is it supposed to be beyond God's power to tell them the truth convincingly enough that they can make a decision based on reality rather than cotton-wool-brained confusion? Or is he supposed to value their ignorance more than their salvation? (A ridiculous idea, for sure, but much of what Christians say about "free will" in these discussions seems to amount to obfuscated versions of that idea.) I think ultimately the answer is that in order to tell the sort of story he wanted to tell, CSL had to make these people ignorant and confused, and greatly limit God's ability or willingness to interact with them, and generally slant things in ways that make good sense narratively but seem awfully hard to justify as things a vastly wise and just and benevolent deity would choose.

(Note: I expect it's already obvious, but: The reason why I keep saying things like "traditional forms of Christianity" is that there are plenty of people nowadays (and there were way fewer in the past) who describe themselves, I think perfectly reasonably, as Christians, but reject some doctrines that used to be taken more or less for granted. The word "traditional" is not intended to express any value judgement, either positive or negative.)

22. lavalamp says:

I think we're all on the same page, but just to be clear: P(C) P(non-christian T|evil) depending on how each treats evil. Christianity is subset of generic theism, so its probability can never be higher than generic theism...

The universe seems much more compatible with an indifferent or malicious deity, so I too am puzzled that very few believe in one. Although, I think a lot of christians (not you, it sounds like) do actually believe in a malicious deity, they just haven't really thought about it carefully enough to notice... :)

23. lavalamp says:

Oh dear, it ate everything between my less than/greater than signs. Oops...

24. Aron Wall says:

Traditional versions of Christianity imply something like the following claim: The majority (perhaps the overwhelming majority) of the human race, if God behaves optimally towards them, will most likely end up in eternal torment and deserve it.

Well, I don't endorse this claim, certainly not as something which I know for sure. For all I know the overwhelming majority of people are going to end up being saved. If I believed that explicit knowledge of Jesus prior to death was necessary for salvation, then this claim would follow, but I don't believe this. (My reasons sre stated on my website in this conversation about why God doesn't speak more clearly.)

For instance, the denizens of CSL's kinda-sorta-pre-hell seem to be almost entirely oblivious of where they are and what their likely fate is. Is it supposed to be beyond God's power to tell them the truth convincingly enough that they can make a decision based on reality rather than cotton-wool-brained confusion? Or is he supposed to value their ignorance more than their salvation? (A ridiculous idea, for sure, but much of what Christians say about "free will" in these discussions seems to amount to obfuscated versions of that idea.)

There is such a thing as the "will to stupidity". Take, as an example, the liberal clergyman character in the book. He is (or was) an intelligent and sophisticated individual who for various reasons dislikes the idea of factual truth as something applicable to matters of religion, or as applied to his current state in existence for that matter. St. Lewis' point is that this is really a disguised version of hatred for God. If he did come to know God, he wouldn't like it, and that's why he remains below.

But one could, alternatively, conceive of Hell as a place where people do perceieve God and the truth with perfect clarity, and find it extremely uncomfortable given their choice of preferences...I'm convinced that if the only problem with someone is that they need more clarity, then more clarity will be provided eventually.

But yes, if Christianity is true then it has to be in some sense an unavoidable a priori feature of reality (based purely on God's nature, logical necessity, and whatever are the basic goods which God wanted to include in the world) that sufficiently free people operating under conditions where God has made himself perfectly obvious may well decide they don't like goodness very much. Angels fall, men fall, the Israelites try to stone Moses, Judas betrays Christ, and God dies.

I don't deny that it's a queer thing. I do feel that some of its psychological aspects are very true to the reality about human beings as I know it (and I don't have a strong sense that I know better than God about the merits of people with radically different kinds of psychology). But this sort of feeling is difficult to communicate (akin to the considerations in Chesterton's book Orthodoxy), and in this blog I'd rather focus on more objective considerations which can be more easily analyzed. Namely the Historical Argument for Christianity, and later (once I've discussed some of the underlying Science a bit more) the Fine-Tuning Argument for Theism.

25. g says:

Aron,

If you think it likely that almost all will in the end be saved, then I kinda approve but I have to point out that this seems hard to square with Matthew 7:13-14, and with the great majority of the Christian tradition until very recently. Any of which you may or may not care about -- but note that if you aim to get out of the antecedent improbability of Christianity (on account merely of being very specific) by an appeal to testimony, then substantial divergence between your views and that testimony is a problem.

The "will to stupidity" that CSL puts in, e.g., the liberal clergyman is one of the more unattractive features of TGD. (I agree that the feature is there; the only reason why I didn't whinge about it before is that I felt I was already writing too much.) It's just one example of the (also unattractive and, I think, grossly unreasonable) tendency of Christians to diagnose people who aren't Christians, or who aren't their sort of Christian, as really being motivated by a "disguised hatred of God". Just consider for a moment how CSL's liberal clergyman got the way he is. How many people choose the life of a clergyman out of hatred for God? I'm thinking not very many. So, he (hypothetically) fell among postmodernists or something; either his condition is curable or it isn't. If it is, then it seems like surely God could have found better means to cure it. If not, then what this amounts to is that in this wonderful world God has made, it is possible -- and not terribly uncommon -- even for those who have chosen a life of dedication to God to talk with the wrong people, contract some diseased modes of thinking, and thereby put themselves into a state from which there is no ultimate way out but eternal damnation. You may decide for yourself what to think of that possibility, but it seems pretty implausible given the characteristics God is supposed to have.

The idea that it's an unavoidable a priori feature of reality that if creatures are sufficiently free then many of them will decide that they don't like God, especially when combined with the idea that God is supremely good and excellent, seems to me not so much "a queer thing" as "absolutely ridiculous". You may well be right that human beings as presently constituted are liable not to like God very much even with perfect information -- but the nature of human beings as presently constituted is precisely one of the things that a skeptic making an Argument From Evil sees as evidence against theism in general and Christianity in particular.

26. Larry Wall says:

g sez:

Larry -- the proposition "the world is, on the whole, a really bad thing, worse than outright nonexistence" is quite different from the proposition "the world is much worse than should be expected if it's the creation of a supremely good and supremely powerful being". The second of those is what's claimed by people making an Argument From Evil; the first is the one you're giving evidence that atheists don't agree with.

You have accurately restated part of my "evidence", and ignored (perhaps unintentionally) the important part, which is the observation that many folks choose, for what I deem to be emotional reasons, to maintain an artificial separation between those two propositions by disingenuously redefining the terms "good" and "powerful", such that "good" becomes the quality of some kind of cosmic Santa Claus who always gives me the toys I think he should, and "powerful" becomes the capability of climbing down as many philosophical chimneys simultaneously as I care to erect. (But perhaps I malign the character of Santa Claus, who is reputed to be capable of putting coal in the stockings of some people who think they deserve toys. :-)

So my main point, which I hope you will evaluate for more than its propositional value, is that, for largely unconscious reasons, some people hold God to a standard they wouldn't hold any human artist to. Note, I am not claiming that religionists are any less emotional about the subject than anti-religionists; if modern psychology has proven anything, it's that we all bring our subconscious biases to the table of our supposed logic, often at the very point where we are defining our table in the first place. Moreover, our personal aversions often influence us just as strongly as our desires, if not more so. It's a very human failing, and tragically/comically, sometimes happens to the very people who most think they are immune to emotional bias. That's been my experience, anyway...

In any case, if any of our Gentle Readers finds this message irritating, please take it not as an accusation but rather as a call for a bit of self-examination outside the scope of propositional logic.

27. g says:

Larry -- I certainly didn't mean to ignore anything relevant. But on what grounds do you say that the separation between those propositions is artificial, and whyever do you think that separating them requires that one define "good" and "powerful" in those extremely-straw-mannish ways?

It seems to me that it's quite right and proper to hold God to a standard you wouldn't hold a merely human artist to, for exactly the same reasons as I would hold a world-class pianist to a higher standard than I do my six-year-old daughter.

I am quite sure that some people -- myself very much included -- make all kinds of errors when thinking about God and evil, for all kinds of reasons (including but not limited to unconscious God-hatred or whatever exactly you're proposing here). For sure, this means we should all engage in self-examination. But what does it mean for the discussion at hand?

Perhaps the idea is that we shouldn't discuss it at all because we might be subject to such biases. I would find it more convincing when Christians make such suggestions when atheists talk about evil, if they made similar suggestions when Christians talked about ... well, anything, actually.

Perhaps the idea is that there is some specific error or class of errors that some of us are making. In that case, let's talk about those errors and try to fix them.

Or perhaps something else?