Home » Library » Modern Library » Bruce Russell Dialogue

Bruce Russell Dialogue


Volume XVI, No. 2, Fall 1988

The “Inductive” Argument From Evil: A Dialogue*


Wayne State University


Calvin College


The following is a conversation between Beatrice Leaver, Athea Ist, and Agnes Tic. Bea and Athea teach philosophy–Bea at a Christian college, and Athea, at a large midwestern university. Agnes is an attorney. Now in their thirties, the women were suitemates in college, together with Iris.

Iris has just lost her six-year old daughter. Bea and Athea have flown to Flint to be with her. They stay with Agnes, who has remained closest to Iris. It is early evening, the day after the burial. Iris, severely depressed, has been hospitalized under medication. Bea and Athea are talking with Agnes in her house.


1. 1 In Agnes’s living room.

Athea: I still don’t understand how it happened, Agnes. Could you go over it again?

Agnes: Well, here are the essentials. Iris divorced Bo eight months ago, and after Iris met Jim, Bo went over the edge. New Year’s Eve, Iris was at the local pub with Jim. Bo came in, and got so abusive he was thrown out. When Jim dropped Iris off, Bo was waiting inside. He threatened her, but Iris took a swing at him and somehow knocked him out. She left him on the floor and went to bed. At 3:45 the brother came in and found Carrie’s body. She’d been beaten over most of her body and strangled to death. Bo was arrested, and Iris thinks he did it. She thinks he tried angel dust again that night.

Athea: It’s all so hard to believe. How is Iris taking it?

Agnes: I’m afraid for her, Athea. She blames herself for not calling the police while Bo was unconscious. She kept saying she couldn’t go on living if she didn’t believe in a loving God–that “all things work together for good to those who love God.” This, for good? To me it is blackness. I keep thinking of Carrie’s innocent eyes . . . and I don’t know whether Iris can recover again. You know how Carrie’s been the center of her life ever since her first husband died.

And I can’t get rid of the thought that this just shouldn’t happen. Remember how, when we were seniors, I stopped believing in God? Well, now and then I’ve wondered if maybe the folks at home weren’t right after all; last fall I even tried going to church again. But this thing–I can’t see how a God could allow it. Of course, I wouldn’t say that to Iris. But I’ve wanted to talk about it with someone. I even dug out my old philosophy book and read an article by an Oxford philosopher named Mackie. Now it hits home.

Bea and Athea (simultaneously): What do you mean?

Agnes (with a wry smile): I’d almost forgotten both of you teach philos­ophy. Remember our ethics prof? “What exactly do you mean?”, he’d ask. Carrie’s death makes me think maybe he was right; he claimed that evils like this conclusively prove that there’s no God–none like the Baptists believe in, anyhow.

[Athea and Bea are silent.]

Agnes: I know it seems weird, but last night I actually worked out an argument, hoping we could talk like we used to in college. Maybe it’s because, not knowing what I believe, I had no words to offer Iris; I could only sit and cry with her. Or maybe I’m using intellectual analysis as a sedative, to escape the pain of these past days.

Bea: Agnes, I don’t think there are words for a time like this. Saint Paul says “Weep with those who weep”; perhaps your tears were far better than any words could be. But if you think some intellectual analysis will be a respite–well, there are much worse ways of escaping.

Athea (staring into her glass of wine): That’s for sure. Maybe talking will keep us from drowning in this stuff. Let’s hear your proof, Agnes.

Agnes: Okay. I started with two simple ideas. The first is that if God is all-good, as Christians say, he has to be against intense suffering, just as a good mother is against her child’s suffering. I mean, a good mother does sometimes allow her child to suffer greatly–say, by having a tonsilectomy. But she doesn’t like this suffering, and she’s not indifferent about it, either; she’s against it, in itself. So she would allow it only if this served some purpose–only if she thought this served some sufficient good. And isn’t that also true for a good God? To allow such suffering for no purpose would mean God is either indifferent to it, or actually likes it for its own sake. And that would mean he’s not wholly good.

Then I tried to clarify what it is to “serve some sufficient good.” I saw that it’s not enough for allowing the suffering to produce the good. Some painful surgery might produce some good result–but suppose there were some way to get this good equally well, but without the suffering. Some “Plan B,” I’ll call it. Wouldn’t a morally good person have to use the Plan B? And isn’t the same true of God, if he is good? I think so. So I wrote:


If God exists, then (being all-good) he would not allow any instance of intense suffering unless doing so served some sufficient good, and there were no “Plan B”–no equally good way to get this good without such suffering.

From there it was simple logic. God is supposed to be omnipotent, so he can do anything. But then he would always have a “Plan B.” It follows, doesn’t it, that if God exists, he would not allow any instance of suffering, so no suffering would occur. But it does occur. So it follows that God doesn’t exist. It looks like I should give up my indecisive agnosticism and embrace atheism.

Athea: You’ve certainly remembered your logic lessons well, Agnes. But there is a problem with one of your premises. Theism does hold that God is omnipotent, but as some theists explain it, this means that he can do anything power can do. And as Aquinas said, power cannot do self-contradictory things, like make square circles. So even though the theistic God is omnipotent, he might, to obtain certain goods, have to allow certain evils. That’s the idea behind the “Free Will Defense”–to have creatures with real moral freedom, God has to allow the possibility of the creature choosing evil.

Agnes: I see what you mean. But Athea, you believe that there is no God. Do you have some better argument against theism?

Athea: I think so. But it’s nearly seven; let’s continue over dinner. Shall we go to the Indian restaurant again?

1.2 At the House of India

Agnes: So what is your reasoning, Athea?

Athea: Like you, Agnes, I believe that a good God would not allow evils like Carrie’s death unless doing so served some “sufficient good,” as you called it. And I, too, tried to prove there couldn’t be any such good. But now I think that’s the wrong approach. The important thing is the obvious thing: no matter how hard we look, we don’t see any sufficient good. This itself gives us good reason to believe there is no such good–and hence, no God.

Agnes: But isn’t that the “fallacy” of arguing from ignorance? Isn’t it like concluding that there is no life on other planets, since we don’t know of any?

Athea: No, I’m not arguing from our ignorance but from our knowledge. We know of particular evils like Carrie’s murder. We also know of many good things, and when we reflect on them, we see that none of them is “God-justifying” with respect to this particular evil. This knowledge that all the goods we know of are “non-God-justifying” gives us good reason to think that all goods are non-God-justifying. This is an inductive inference of the sort we rely on all the time. For example, that we know of no copper that is insulative gives us good reason to believe that no copper is insulative. That we see no elephant in this room gives us good reason to think there is no elephant in the room. These aren’t arguments from ignorance; neither is mine.

Agnes: So your argument isn’t meant as proof giving 100% certainty, but as inductive evidence justifying a high degree of confidence. Is that what you mean by “good reason”?

Athea: Exactly. And because my evidence doesn’t give 100% certainty, it could be outweighed if there were enough evidence that God does exist. But I don’t think there is. So on the basis of my inductive evidence, I think reason requires us to believe that there is no God-justifying purpose for certain evils, and hence, that God does not exist.

Agnes: That sounds plausible to me. But our philosophy prof always said we should test arguments by seeing if they can survive serious objections. So let’s do that with yours. It would be nice if we could write it out premise by premise. Unfortunately we don’t have any paper . . .

Athea: No problem–philosophers write on whatever is at hand, so I’ll just use this napkin. My central argument is this:

(P1) Carrie’s murder was, in itself, a very bad thing.

(P2) If there was no sufficiently good point served by allowing this very bad thing to happen, then God, if he exists, would have prevented it.

(P3) There was no sufficiently good point served by allowing this very bad thing.

(C1) Therefore, if God exists, he would have prevented this thing from happening.

(P4) God did not prevent this thing from happening (since it did happen).

(C2) Therefore, God does not exist.

Now premises 1 and 4 seem beyond dispute. And we’ve already agreed on premise 2 in discussing your argument, Agnes. So the only premise to worry about is P3. And the heart of my case is that we can defend P3 inductively. Since this will be a subargument for P3, I’ll write it using lower case indices. Bea, could you pass that napkin?

(p1) After careful reflection, we see no good point served by allowing this bad thing to occur.

(p2) If, after careful reflection, we see no sufficiently good point served by allowing some bad thing to occur, then we have some reason to believe there is no sufficiently good point served by allowing it to occur.

(c1) Therefore, we have some reason to believe there is no sufficiently good point served by allowing it to occur.

(p3)We have no outweighing reason to believe the contrary (that there is a sufficiently good point served by allowing this evil).

(p4) If we have some reason to believe some proposition, and no decent reasons to believe the contrary, then on balance, reason requires us to believe that proposition.

(c2) Therefore, reason requires us to believe that there is no sufficiently good point served by allowing this bad thing to occur.

[Athea slides the napkin to Bea and Agnes]: Voila!

Agnes: I see what you’re doing. You first conclude, at c1, that your inductive evidence gives some reason to believe P3 of your main argument. You then infer, at c2, that since there is no evidence to outweigh this reason, we are rationally required to believe P3. Bea, what do you think about this?

Bea: Interesting–by stating her inductive principles as premises p2 and p4, the subargument for P3 has become a deductively valid argument. As you know, I believe in God. Since I reject the conclusion, I’ve got to look for some false premise, and I think the problem must lie in this subargument from our not seeing any God-justifying good. It reminds of those bugs we have here: though you feel their bite, they are so small you can’t see ’em, so they’re called “noseeums.” I’d like to baptize Athea’s subargument the “noseeum argument.” And I see three things to scrutinize in it.

First, we might scrutinize p1. Don’t we see any good that might justify God in allowing this evil? What about free will? But suppose that doesn’t work, and we see no other sufficient good. We must then look carefully at p2. Is seeing no good really evidence that there is no such good? Perhaps this “inductive” inference is fallacious in some way. If neither of these work, we might challenge p3. Even if Athea has given us some evidence there is no such good, can’t we find positive evidence outweighing this? Evidence that God exists would be evidence that there is some such good even though we don’t see it. Are we so sure there isn’t evidence for God sufficient to balance Athea’s negative evidence?

Agnes: Well, I want to discuss each of them, even if it takes all night. When else will I be able to talk with two philosophers like you? But why don’t we pay the tab and go down the street–there’s a place called “Mother’s” that has absolutely sinful desserts.

1.3 At Mother’s

Agnes: Let’s begin with p1, then. Bea, what good might justify God’s allowing a thing like Carrie’s murder? As a Christian, surely you think you see something explaining it.

Bea: Be careful with the “surely” there, Agnes. Christians don’t claim they are omniscient. But some theists do propose some possible goods that might do the job. In the case of a moral evil, like this brutal murder, human choice is clearly involved. On one view, God sets before us options for good and evil, and, to some extent, leaves it up to us which we choose.

He has placed us in a world in which our choices make real differences, not just to ourselves but to others: whether others flourish or perish is, in many ways, up to us. Isn’t it a good thing that we have this kind of significant freedom, and a world in which it can be meaningfully exercised?

Agnes: I suppose it is a good thing. And it does seem that this freedom would be compromised if God always stopped moral evils from occurring. Suppose he made bullets vaporize before they struck humans–or made everyone like superwoman, with no kryptonite around. Then our world would be one in which we could not choose to hurt another. Perhaps we could form intentions to hurt others, but we’d soon learn we couldn’t ever carry them out. But then we’d lack freedom. So if God wants a world with this good of freedom in it, doesn’t he clearly have to allow evils resulting from free choices, including the evil of Carrie’s being murdered? What do you say to that, Athea?

Athea: Even if I grant that moral freedom is a good thing, there are problems. Couldn’t an omnipotent God make us so we are free, but always freely choose to do right? This doesn’t seem a contradiction–some Christians even believe that angels like Gabriel are like this. Couldn’t Omnipotence make us all like Gabriel? If not, is freedom to hurt others really worth the price tag? Like Dostoevsky’s “Grand Inquisitor,” I’m not sure such freedom is worth the torment of one little child like Carrie.

If these problems are solved, the biggest problem remains. For it would just mean that God, to have such freedom in his universe, must sometimes allow moral evils. Not that he’d always have to allow them. A mother, to teach her daughter responsibility, must sometimes allow what results from her wrong choices. But she needn’t always do this: she could still intervene sometimes to prevent dreadful catastrophes. Similarly, obtaining moral freedom doesn’t require God always to allow moral evils. So this good doesn’t clearly explain, or justify, his allowing the moral evil of Carrie’s being murdered.

Bea: I tend to agree. Theists, unlike eighteenth-century deists, believe God is active in his creation. He’s not a clockmaker who made the world and left it to run on its own. As a theist, I think that God does sometimes intervene to prevent moral evils. It is only his doing this always that would destroy freedom.

Agnes: So Athea, how does this help you? We’ve agreed that if God exists, he will sometimes intervene to reduce evil. So what? In concluding that God doesn’t exist, are you supposing that such interventions never occur? What evidence is there for that? Maybe Bo was first going to kill the whole family, and God intervened by helping Iris to knock him out. Bo is no midget, you know.

Bea: No, Agnes; remember how this issue arose. The question was whether we see any good that would justify God’s allowing this evil. You suggested that to have the good of freedom, God clearly had to allow Carrie’s murder. This, Athea is saying, falsely assumes that our having freedom requires no interventions from God. For we’ve agreed that God can sometimes intervene without destroying this good of freedom.

Athea: Thanks, Bea. I see why your students are never sure whose side you are on. But you’re exactly right. To see moral freedom as a good justifying God’s allowing Carrie’s murder, we would have to see that his intervention on that occasion would have compromised this good. My argument is that we don’t see that at all. In fact, I think (though my argument doesn’t demand this) that we see the contrary. If Iris’s knocking Bo out the first time (with or without Divine assistance) didn’t compromise his freedom, it sure seems that stopping him one more time wouldn’t either.

Agnes: Wait a minute. I just remembered a principle from my ethics class. It was that similar cases must be judged similarly. If, under certain circumstances, I should keep a promise, then under similar circumstances, you should too. If one sees no relevant differences in the circumstances, fairness requires there be no difference in one’s judgment about what should be done. So if we accept your claim that God should intervene to prevent Bo from carrying out his murderous intention on this occasion, won’t you have to say that about the next occasion too? And so also for the next occasion. But that would mean he should always so intervene, which you granted is false, for that would destroy freedom. Doesn’t this refute your claim? Or do you reject the principle about judging similar cases similarly?

Athea: No, I don’t reject the principle: if the circumstances aren’t different, the judgment shouldn’t be either. But we need a broad enough notion of “circumstances.” If drunk drivers are killing people, it might be right for the police to test randomly selected motorists at checkpoints, but not right to test everyone. One might think that this violates your principle: isn’t stopping one like stopping another? Not really. There might be a certain threshold number of people stopped, such that as you pass that threshold you start getting bad effects–say, causing long delays, so that drivers get hostile and cause more accidents than drunkards. The number of drivers you’ve already checked can be part of the circumstances in deciding whether to check one more driver. So you might rightly not stop one more driver, because as you reach that threshold, the circumstances are different.

So also with divine interventions. It might be right for God (if he existed) sometimes to prevent a person’s evil actions, but after a certain number of interventions, a threshold may be reached where further interventions would have very bad effects-say, the person feeling that he lacks real freedom of action. As that “freedom threshold” is reached, it might not be right for God to intervene once more.

Bea: That seems sensible to me, Athea. I would also apply it to God’s intervening to prevent “natural evils,” due to accidents and the like. The real problem now seems to be this. You allow that to have the good of moral freedom, there are “freedom thresholds” which, once reached, would justify God’s allowing moral evils of murder and the like. But you don’t think this good justifies God’s allowing this evil of Carrie’s death. So you must be supposing the “freedom threshold” has not been reached in this case. What reason is there to suppose that?

Agnes: Can I suggest one? Suppose you were a cop and you saw the little girl being brutalized by someone. Would you have thought: “Well, I’ve got to be careful: maybe one more intervention will cross some threshold, making the person feel he lacks choice, so he doesn’t try to be good any more than I try to fly?” Of course not. If anyone proposed this as an excuse for your not intervening, you’d dismiss it as ludicrous. Why should we take it more seriously as an excuse for God’s not intervening?

Bea: I once heard a similar argument from a philosopher named David Conway. At first it seemed quite compelling. But then I read The Existence of God by Richard Swinburne. Swinburne criticizes the assumption that since we’d have a duty to stop something, so would God. There is, he says, no reason to think God is in the same position as we are. For one thing, if God exists he knows much more than we do; and differences in knowledge can make differences in duties. If someone is choking, it might be right for a doctor to cut open his throat, but wrong for me (lacking his medical knowledge or skill) to do so.

Athea: I agree, Bea, that one can’t simply argue “from us to God.” But there’s a better way to interpret Conway’s argument. What is important is why we think the cop should intervene: we judge that his intervening would not cross any “freedom threshold, ” that is, would not begin to compromise freedom in a way that produced threshold effects worse than the evil at issue. There may be some such freedom threshold, but we are confident it is not here, or we would not intervene so confidently ourselves.

Here’s another idea. Why do we think that one more human intervention would not cross any “freedom threshold?” I think it’s because we don’t see how it would. So we are relying on a “noseeum” inference. Well, we also don’t see how a further divine intervention would cross a “freedom threshold.” So isn’t a noseeum inference justified here too? Especially since God could intervene covertly, say, by waking Iris with the thought of checking the children. If the freedom threshold Wouldn’t have been violated by Iris’s waking naturally and stopping Bo, why would it by God’s waking her?

Bea: I don’t know, Athea. I think you’ve put your finger on an important problem, and I don’t know how to answer it. It strikes me that the problem also arises for “natural evils,” and that even believers can be gripped by it.

Let me explain. We’ve seen that if God always intervened to prevent evils, certain goods might be lost. But theists believe God does often guide and protect us. Perhaps we’ve had loved ones escape tragedy by some fortunate coincidence: believers (if they are not deists) will see this as God’s providential care. But what, then, if my child drowns, and wouldn’t have, if the lifeguard had only turned his eyes to the corner of the pool? God averted one close call last month; couldn’t he avert one more now? It seems so implausible, especially when tragedy strikes close to home. You’ve helped me see an important problem, and I want to think about it more.

Agnes: It makes me wonder if deism isn’t the best theism after all. Your God intervenes to protect little children sometimes, but other times he must regretfully let nature take its toll on them. How does he decide? “Shall I protect this child? Careful now–I’ve already had to protect X children already this month.” I can imagine God thinking: “The heck with this; it’s too much fiddling around. I’ll just make a universe and let it roll.”

Bea: I can feel the pressure toward that way of thinking, though I think it’s to be resisted. But for now, I will concede Athea’s first premise. We do not see a good justifying God’s allowing Carrie’s being murdered. Maybe free will is a good justifying God’s allowing this; but given Athea’s threshold arguments, I don’t see how it does so. So why don’t we turn to p2 in her sub-argument? Suppose we don’t see any God-justifying good for this. Is this evidence there is no such good?

Agnes: Why don’t we continue over a brandy? There’s a quiet place, the Barrister, just around the corner.


2.1 At the Barrister: Round One.

Agnes: Now that we’ve got our brandies, let’s turn to Athea’s p2: “If, after careful reflection, we see no sufficiently good point served by allowing some bad thing to occur, then we have some reason to believe there is no sufficiently good point served by allowing it to occur.” Bea, what do you have against this?

Bea: My objection has two points. First, suppose we took p2 as a general rule: our not seeing something after careful looking gives us reason to believe it’s not there. Let’s call this a “noseeum rule.” It seems clear that as a general rule it is false: it works for some cases, but not for others. Seeing no elephant in a normal room, after looking hard, gives us good reason to believe no elephant is in the room. But suppose we wonder if a sandflea is in the room. Looking hard, we see none. Is this good reason to think none is in the room? Clearly, far less so than in the elephant case. How about a strep virus? Does seeing no strep virus give us a reason to think none is in the room? Not any reason worth mentioning, surely. So clearly, the acceptability of a noseeum rule depends on the kind of critter at issue. Let’s call the following my “expectability principle.” If the critter has low “seeability”–if it is the kind of critter that, under the circumstances, you wouldn’t expect to see even if it is there–then the noseeum rule is false: not seeing it is not evidence it’s not there.

My second point is this. We want to know if Atthea’s noseeum data is evidence that God doesn’t exist. At issue is whether there is some good purposed by God, justifying his allowing this murder–some “God-purposed good,” let’s call it. Does our seeing no such critter give us reason to think none is there? If my first point is correct, this depends upon the “seeability” of such a critter: if there is a God-purposed good for this, would it likely fall within our ken? I think we have good reason to answer “No.” A God-purposed good would be more like a strep virus than an elephant. Given my first point, then, the noseeum rule is false for such goods.

Agnes: You say we have good reasons to think that if there is a God­purposed good for this, it would likely be beyond our ken. What are these good reasons?

Bea: Well, I just read one reason in the 1984 issue of the International Journal for Philosophy of Religion, which I happen to have with me. Here it is, on page 88, by a Stephen Wykstra:


We must note, first, that the outweighing good at issue is of a special sort: one purposed by the Creator of all that is, whose vision and wisdom are somewhat greater than ours. How much greater? A modest proposal might be that his wisdom is to ours, roughly as an adult human being’s is to a one-month old infant’s. (You may adjust the ages and species to fit your own estimate of how close our knowledge is to omniscience.) If such goods as this do exist, it might not be unlikely that we should discern some of them: even a one-month old infant can perhaps discern, in its inarticulate way, some of the purposes of his mother in her dealings with him. But if the outweighing goods of the sort at issue exist in connection with instances of suffering, that we should discern most of them seems about as likely as that a one-month old infant should discern most of his parents’ purposes for those pains which they allow him to suffer. Which is to say, it is not likely at all. So for any selected instance of intense suffering, there is good reason to think that if there is an outweighing good of the sort at issue, we would not have epistemic access to this . .

Athea: Pass that here, Bea. I know this article: Wykstra is criticizing a “noseeum” argument by William Rowe. But as I recall, it is followed by a reply from Rowe that demolishes Wykstra’s argument. Here it is on page 97:


It’s true, as Wykstra observes, that God’s mind can grasp goods that are beyond our ken. The idea, then, is that since God grasps goods beyond our ken, we’ve reason to think it likely that the goods in relation to which God permits many sufferings that occur would be unknown to us. Let’s look at Wykstra’s reasoning here. He starts with:

1. God’s mind grasps goods beyond our ken.

moves to:

2. It is likely that the goods in relation to which God permits sufferings are beyond our ken.

and concludes with:

3. It is likely that many of the sufferings in our world do not appear to have a point–we can’t see what goods justify God in permitting them.

The difficulty is that the move from (1) to (2) presupposes that the goods in question have not occurred, or, at the very least, that if they have occurred they, nevertheless, remain quite unknown to us (in themselves or in their connections with the suffering in our world). And, so far as I can see, the mere assumption that God exists gives us no reason to think that either of these is true. If God exists it is indeed likely, if not certain, that God’s mind grasps many good states of affairs which do not obtain and which, prior to their obtaining, are such that we are simply unable to think of or imagine them. That much is reasonably clear. But the mere assumption that God exists gives us no reason whatever to suppose either that the greater goods in virtue of which he permits most sufferings are goods that come into existence far in the future of the sufferings we are aware of, or that once they do obtain we continue to be ignorant of them and their relation to the sufferings.

So Rowe concludes that Wykstra has not justified the claim that if God were to exist, the sufferings in our world would appear to us as they do­-that is, that we would see no point to them. Wykstra’s argument thus fails.

Bea: But has Rowe correctly represented Wykstra’s argument? It seems to me Rowe has missed something–though perhaps Wykstra didn’t spell it out sufficiently. Rowe takes Wykstra to start from the true claim that God, if he exists, sees goods we don’t see; Rowe then objects that this fails to show that God allows present sufferings for the sake of goods we don’t see. But I don’t think Wykstra means to appeal just to God’s superior knowledge. Look again at how he opens: the good at issue would be “of a special sort,” he says, being one “purposed by the Creator of all that is.” This, ignored by Rowe, makes a big difference.

Agnes: How does it make a difference?

Bea: We can see it this way. First imagine a being who is good, omniscient, and has great powers of intervention, but who is not a Creator. He just “came upon” our universe as a brute fact, the way we did, and is hanging around being good. The mere fact that this being sees many goods we don’t see–goods thousands of years in the future, for instance–may have little bearing on whether, in allowing some current suffering, this being has in mind one of these goods that are distant or otherwise unknown to us. On this Rowe seems to me to be right. But now add the fact that Rowe ignores: that God is Creator of all that is. If he exists, then as Job is reminded, he laid the very foundations of our universe: it proceeded from his goodness and wisdom. Adding this makes a big difference, because it raises a whole new question: if our universe is the creation of such a Being, whose wisdom is to ours as an adult’s is to a one-month infant’s, what should we expect on the present issue?

Agnes: I see that it makes a difference. But how big is it?

Bea: Let me get at that indirectly. We today know something about physical explanations–explanations in terms of electrons, protons, and the like. For most of history, humans didn’t even have the concept of such explanation. Around 1600, Francis Bacon had some concept of it, but he thought that if men would only apply his experimental method, physics could, in about twenty years, get to the bottom of what causes what. He saw our universe as being, we might put it, “physically shallow–as having bottom-line physical causes that are, relative to our cognitive abilities, quite near the observable surface. Newton, comparing his greatest discoveries to picking up shells on the ocean’s shore, had a contrary view. Four hundred years of science have vindicated Newton. Having descended into a swarm of quarks, leptons, and other denizens of the microtheoretic deep, we are more astonished than ever. The “bottom-line” micro-causes of the observable world lie very deep indeed. We realize our world has extraordinary “physical depth.”

Now I want to introduce a similar concept–that of “moral depth.” If God exists, then as Athea says, he would allow an evil only if doing so served some outweighing good. Such a good would be the “moral (or axiological) cause” for his allowing the evil. If God exists, there is some such cause for each of his allowings of evil. But here there are two options. We might think that if our universe is the creation of God, these moral causes would likely be “near the observable surface” of their effects. This would be a “Baconian” view of the universe as morally shallow. In contrast, we might judge that if our universe is the creation of God, it would likely have great “moral depth.” By this I mean that many of the goods below its puzzling, observable surface, many of the moral causes of God’s current allowings and intervenings, would be “deep” moral goods. The question is this: if our universe is the creation of God, is it more likely to be morally shallow or morally deep? What Wykstra means to say, I think, is that if our universe is the creation of God, a God with the sort of wisdom and vision entailed by theism, then it is eminently likely that it is morally deep rather than morally shallow. He is not, as Rowe supposes, appealing merely to the fact that God knows more than us. He is appealing to the fact that if this sort of God does exist, the axiological foundations of our world were laid by His wisdom and vision.

Agnes: Unless you can read Wykstra’s mind, that’s a lot to get out of Wykstra’s opening words. But I see what you mean, and it does seem to pose a new challenge for Rowe. Athea, I’d like to hear what you think? How about another round of brandy?

2.2 At the Barrister: Round Two

Athea: I have a few worries about the claim that if God exists, our universe would likely have moral depth. I take it, Bea, that you mean to define a “morally deep” good so that it would likely be inscrutable to us. But if God is good, and cares about us, wouldn’t he want us to be apprised of his game plan? Wouldn’t he want the universe to be morally transparent (I prefer this term to “morally shallow”) to sensitive creatures like ourselves?

But my main objections to Bea’s moral depth line are different than Rowe’s. I’m willing to grant, at least for now, that if God exists, it may be quite likely that our universe would be morally deep (or, as I’d call it, “morally obscure”). But even granting this, I think your case fails for very different reasons. Here I should acknowledge that I owe these objections to Bruce Russell, a close friend of mine. I think they are on target, and in letters, even Rowe seems to like them.

The first objection is this. To make your case using your expectability principle, you need to show that it is reasonable to believe that if this tragedy has a point (or serves some outweighing good giving it a point), then we would likely not see this point. But you have not shown this. All you have argued is that if God exists, then if (though) the suffering has a point, we likely would not see it anyway. But to show what, according to your principle, you need to show, you must now establish the antecedent-that God does exist. Otherwise you beg the question, by assuming that God does exist. Of course, I know that in this context, you don’t want to have to show that God exists. In that case, what you really need to show is that it is reasonable to believe that whether God exists or not, if this tragedy serves some outweighing good, we likely would not see it. You haven’t done this; that’s my first objection.

Bea: There is something that confuses me here, Athea. What you grant I’ve shown is, it seems to me, all I need to show. But suppose we hear your second objection before examining this.

Athea: My second objection is that your expectability principle is false. I shall give what I think is a decisive counterexample to it. To begin, I should introduce a principle that my counterexample rests on. It is this: if, given some evidence e, I am justified in believing some proposition p, and I know that p entails q, then given e, I am also justified in believing q. Call this my “transmission principle.”

Agnes: That seems right. Let’s hear your counterexample.

Alhea: Well, consider the fact that I now have the visual sensation of there being a table in front of me. Call this my “visual data.” Unless we are skeptics, I take it that we agree that given this visual data, I am justified in believing that I see (really see) a (real) table in front of me.

Agnes: Didn’t Descartes have some nasty arguments about that–didn’t he propose that there might be an “evil demon” who is bent on deceiving me, causing me to have sensations of physical objects that aren’t there?

Athea: Exactly–that’s the next element in my counterexample. Consider the hypothesis that there is an evil demon causing me to have sensations of a non-existent table. Call this the “demon table” hypothesis. Now there are two things we must note about this hypothesis. First, we know that my belief that I see a table in front of me entails that this hypothesis is false. If I see a table in front of me, then there is no evil demon causing me to have sensations of a non-existent table. By my transmission principle, we must then say that I am justified, on my visual data, in believing that the demon table hypothesis is false. And this fact is really just common sense.

But if we accept Bea’s expectability principle, we must deny this common sense fact. The demon table hypothesis makes entirely likely (it entails) that I would have just the visual table-data I have, so by her rule, this data cannot be evidence justifying my belief that the demon table hypothesis is false.

Agnes: So you are giving a “Reductio ad Absurdum” of Bea’s principle. From her principle, it follows that on my visual data, I am not justified in believing the demon table hypothesis is false. But I am justified in believing the demon table hypothesis is false (this is the “common sense” point you support by your principle). So her principle must be mistaken.

But surely there is something right in her expectability rule. In the examples she gave–the strep virus and sandflea cases–her rule was very plausible. Can you account for this?

Athea: I think there is a correct principle in the neighborhood of Bea’s, but it does not give the results she wants. The correct principle is rather complicated. Let’s consider an example. Normally our not smelling any sourness in a glass of milk is evidence that it’s not sour. But, of course, if you knew you had a cold, your not smelling any sourness wouldn’t be evidence that it’s not sour. The correct principle explaining this is the following:


If (1) the milk has no sourness that I can discern, and (2) I have reason to believe if I have a cold, the milk would have no sourness I could discern even if the milk were sour, and (3) I have reason to believe I have a cold, then my not discerning any sourness gives me no reason to think the milk is not sour.

More generally, let X and P be the object and property at issue, and C a condition under which P will be “unseeable.” Then:


If (1) X has no P that I can discern, and (2) I have reason to believe if C obtains, X would have no P I could discern even if X were P, and (3) I have reason to believe C obtains, then my not discerning X as having P gives me no reason to think X is not P.

If we apply this principle to a case (X) of tragic suffering which serves no point (P) we can discern, we get:


If (1) the tragedy has no point that I can discern, and (2) I have reason to believe if God exists, we likely would not see the point of the tragedy even though it has a point, and (3) I have reason to believe God exists, then my discerning no point to the tragedy isn’t reason to think it has no point.

Agnes: So you think Bea needs each of the three things in the antecedent, but establishes only the first two. This connects back to your first objection then. doesn’t it?

Athea: Yup.

Agnes: Well, this place is closing. I propose we go down the street to Denny’s, and hear Bea’s response there.

3.1 At Denny’s: Coffee Pot #1

Agnes: What about Athea’s first objection, Bea? She says you show only that if God exists, then if the tragedy has a point, we likely wouldn’t see it. You need to show that if the tragedy has a point, we likely wouldn’t see it. Don’t you slide from one to the other, begging the question by supposing that God exists?

Bea: Whether I’ve begged the question depends on exactly what the question is. Athea adduces her noseeum data as significant evidence that God doesn’t exist. Why? Because if God did exist, he would allow a tragedy like Carrie’s death only if doing so served some purpose, some sufficient good. So I take the real question to be whether Athea’s noseeum data is evidence that there is no such God-purposed good. To establish a “No” answer to this question, all I need to show is that if there were a God­purposed good–one actually purposed by God–which justifies his allowing the tragedy, we likely wouldn’t see it. All I need to show, in other words, is that if God exists, then if the tragedy serves some good, we likely would not see it. And that, Athea allows I have shown.

Agnes: What about Athea’s cold analogy? There, the question is whether smelling no sourness in the milk is evidence that the milk is not sour. Suppose you wanted to argue it isn’t. Surely, as Athea says, you couldn’t merely argue that if I have a cold, then if the milk is sour, I’d likely smell no sourness anyway. To defeat the nosmellum data, you’d have to go on to establish that I do have a cold. Similarly, to defeat Athea’s noseeum data, don’t you have to go on to show that God does exist?

Bea: There is a disanalogy between the two cases. Consider the two properties at issue. In the cold case, the issue is whether the milk has the property of being sour. Now there is nothing about this property itself that makes the nosmellum data expectable. For precisely this reason, defeating the data requires establishing that some further condition obtains (like having a cold), under which, even if the milk has the property, the nosmellum data would be expectable.

In the God case, the issue is whether some suffering has the property of serving some God-purposed good. Here there is something about this putative property itself–serving a God-purposed good–that leads us to expect that if the tragedy has this property, we’d get our noseeum data anyway. For precisely this reason, we don’t need to establish some further condition under which, if the suffering has this property, we’d nevertheless expect the noseeum data.

Agnes: I think I see what you are saying, but Athea’s analogy still bothers me.

Bea: Let’s loosen the grip of her analogy by considering a different one, which doesn’t have the dissimilarity I’m pointing out. Suppose you, an intelligent philosophy undergraduate, find a philosophy manuscript in a dorm room. It has many sentences for which you cannot see any meaning. You are about to dismiss it as nonsense, when it occurs to you that a visiting philosopher might have stayed in the room. And some philosophers, as you know, write papers which are so brilliant and deep that they are intelligible only to a handful of fellow specialists. Perhaps, you conjecture, this manuscript was written by such a philosopher–a Dr. Genius, as you call her.

Agnes: I take it a Dr. Genius does not write gibberish, so if it was written by a Dr. Genius, then these sentences would all have meanings, but these would often be beyond my ken–being “Genius-meanings,” so to speak.

Bea: Exactly. Of course, you don’t know whether it was written by a Dr. Genius; this is just one possible hypothesis. Now your noseeum data is that you see no meaning for many sentences in this manuscript. The question is whether this noseeum data is significant evidence against the hypothesis the manuscript is by a Dr. Genius.

Agnes: I don’t see that it could be. We’ve agreed that if it was by Dr. Genius, then one would expect many sentences to have no meanings I could see. Since the noseeum data is just what we’d expect if the hypothesis were true, how can it be evidence against the hypothesis?

Bea: But suppose someone objected that all you’ve shown is that if the manuscript is the product of a Dr. Genius, then if some sentence has a meaning, we’d expect not to see its meaning? To avoid begging the question, don’t you now have to show it is the product of a Dr. Genius?

Agnes: Of course not. The question is whether the noseeum data is evidence against the claim that some sentence has a meaning-intended-by-Dr.-Genius. To show it isn’t, all I’d have to show is that if it had that sort of meaning, we’d expect not to see it anyway. And that’s the same as showing that if the manuscript was written by Dr. Genius, then if some sentence has a defensible meaning, we’d likely not see it anyway.

Bea: The same for Athea’s argument. To defeat her data, I need only show that if some tragedy has a purpose intended by God, we’d likely not be able to discern it. That’s the same as showing that if God exists, then if the suffering serves some justifying purpose, we’d likely not see it anyway.

Agnes: Now I see why you think her sour milk analogy is irrelevant. There’s nothing about sourness itself that makes you expect the nosmellum data; so to defeat such data, you’ve got to show there is some further condition, like having a cold, which makes this data expectable. But in the God case, your claim is that the very nature of the property at issue­-serving a good intended by God–leads one to expect the noseeum data. But I’m just repeating what you said, aren’t I?

Bea: Yes, but I don’t mind. It doesn’t happen very often.

Agnes: What about Athea’s demon table argument, Bea?

Bea: To begin, we must follow Carnap’s distinction between two senses of “confirms.” Evidence E confirms hypothesis H in a static sense when, given E, H is more likely to be true than false. E confirms H in a dynamic sense when it raises the probability of H higher than it was prior to E. A parallel distinction obviously holds for two senses of “disconfirms.”

Agnes: So let H be Iris’s hypothesis that Bo is the murderer, and suppose we have evidence showing that the murderer is one of Iris’s past boy­friends. Since she’s had a number of boyfriends, this evidence might not confirm H in the static sense. But it might well dynamically confirm the hypothesis, raising its likelihood from what it was before we had this evidence.

Bea: Exactly. Now I’ve put this in terms of probabilities. But a similar distinction holds for justification, I think. E justifies H in a static sense when, on E, one is justified in believing H. E justifies H is a dynamic sense when it increases the justification of holding the belief, or contributes to the belief’s being justified. The key issue is whether Attica’s noseeum evidence dynamically confirms atheism, increasing its likelihood, or contributing to its justification. That’s equivalent to whether it dynamically disconfirms theism, decreasing its likelihood or justification. I argued it doesn’t, relying on the “expectability rule” that insofar as E is “expectable” on some hypothesis, it can’t dynamically disconfirm the hypothesis.

Agnes: Now let me restate Attica’s counterexample. Her visual data justifies her belief that she sees a real table; and this, she knows, entails that it’s not a demon table. So by her “transmission principle,” her visual data also justifies her belief that its not a demon table. But this means your expectability rule is wrong. For if there were a demon table, we would expect just this visual data; so your principle entails, absurdly, that her visual data cannot justify her no-demon-table belief.

Bea: Good job. Now I think this argument equivocates on the term “justifies.” Since her visual data justifies her belief that she sees a table, she concludes, by her transmission principle, that it also justifies her no-demon­table belief, and this, she charges, is absurdly denied by my expectability rule. But I think her transmission principle is true only if we take “justifies” in its staticsense; so it only allows her to conclude that her no-demon­table belief is statically justified on her data. And my principle does not rule that out. My expectability rule only claims that her visual data cannot dynamically justify her no-demon-table belief; it cannot increase its likelihood, or contribute to its justification. Athea’s transmission principle­-and hence her argument–gives no reason to reject that claim.

Agnes: Maybe not; but is the claim plausible?

Bea: I think so. Suppose Athea wakes up one morning next to her precious antique table. With eyes still closed (perhaps she’s been studying too much Descartes), she considers the proposition that an Evil Demon is ready to produce a “demon table” when she wills to open her (demon) eyelids. She believes this hypothesis is false; she rates it as ludicrously improbable; and we, being non-skeptics, deem her justified in this. Now she opens her eyes and has visual table-sensations. We will grant, of course, that on her new visual data, she is still justified in disbelieving the demon-table hypothesis. But why should we see her visual data as increasing this justification, or making less justifiable the demon-table hypothesis?

Agnes: Just one more question. Why do you think her transmission principle is true only in the static sense of “justifies”? If some evidence E increases the likelihood (or justification) of my belief that p, and I know that p entails q, doesn’t E increase the likelihood (or justification) of belief q, too?

Bea: Nope. Take the proposition that Athea is a woman philosophy professor. This is a conjunction of three conjuncts: Athea is a woman AND Athea is a prof AND Athea is a philosopher. Now suppose E is that Athea has published in the American Philosophical Quarterly. E might greatly raise the likelihood of the third conjunct, and hence, might greatly jack up the likelihood of the whole conjunction. But it needn’t raise at all the likelihood of the first conjunct–that Athea is a woman–even though this is entailed by the conjunction. So the transmission principle clearly isn’t true if “justifies” is used in the dynamic sense.

Agnes: Is it true even in the static sense? Couldn’t the conjunction be more likely true than false on some body of evidence, even though one conjunct has a likelihood less than .5 on that evidence? Maybe this conjunct has a likelihood of only .3, but the other two conjuncts have a likelihood of .9; when you average them, the conjunction would have a likelihood of .7, which is more likely true than false.

Bea: No, Agnes, it would be .243. The probability of a conjunction is not the average of the probabilities of its conjuncts, but (assuming they are independent) the product of them. So, like the proverbial chain, a conjunction will never be more likely than its least likely conjunct.

Agnes: Oh.

3.2 Coffee Pot #2

Athea: Well, we’d better have another pot of coffee. Let’s make it decaf, so we can get some sleep tonight.

Bea, you said that the real question was whether this evil serves any “God-purposed” good–meaning a good which is both purposed by God in allowing the evil, and sufficient to justify his allowing it. But I don’t see that as the original question at all. Let’s look at our napkins from House of India. (She shuffles through them until she finds the right one.) Here, look. P3 in the main argument, which my subargument supports, really says that there is no good thing that would suffice to justify God, if he exists, in allowing Carrie’s murder to happen. Such a good–a God­-sufficing good, I shall call it–could exist even if it is not purposed by God, and even if God doesn’t exist. My argument is that our noseeum data is reason to think there is no such God-sufficing good. And you have given no reason to deny that, for you haven’t shown that if there is a God­-sufficing good, then we likely wouldn’t see it anyway.

So it seems to me you are criticizing a different argument than the one I gave. First you foist on me this argument: if God exists, he would purpose some good in allowing Carrie’s death; there is no such God-purposed good; so God doesn’t exist. Then you pretend the question is whether my evidence gives reason to believe there is no God-purposed good. Well, I agree that our seeing no God-purposed good gives us no reason to believe there is none. But this is a red-herring, diverting us from the argument I actually gave. It’s a good thing Agnes had us write it down.

Agnes: So you think Bea’s criticism is beside the point because it’s not directed at the argument you gave. I guess you’d say the same thing about her Dr. Genius analogy. Suppose that if a sentence was written by Dr. Genius, then we likely wouldn’t see the meaning of it. Then Bea is right: our not seeing the meaning of a sentence is no reason to think it wasn’t written by Her Brilliancy. But you’d say that, nevertheless, we can be justified in believing that there is no Dr. Genius whose meaning lies behind a sentence that seems gibberish; so also, we can be justified in believing there is no God whose good lies behind suffering that seems pointless.

Athea: Exactly. Consider the Cartesian demon again. Surely . . .

Bea: Um, could I interrupt? Athea, you called my moral depth line a red herring. I don’t agree, but I see that I didn’t make its relevance clear. You’re quite right that your original argument was not in terms of God-­purposed goods. But I didn’t mean to “foist” that on you. I meant that logically, the real issue is whether your evidence makes it less likely that there is a God-purposed good.

Agnes: But why is this the “real” issue? Athea agrees that her noseeum evidence doesn’t make it less likely that there is a God-purposed good. But in her argument, she intended her evidence to justify the claim there is no God-sufficing good for E. Since we are evaluating her argument, why isn’t the “real” issue whether her evidence justifies that claim?

Bea: Well, that too is the real issue. Here we must be careful: we need to see how the two issues are logically related.

Maybe the Dr. Genius case can clarify this. Let’s say that a sentence has a “defensible meaning” when it has a meaning on which the sentence is not patently false; and suppose that genius doctors don’t write patently false sentences. So if a sentence is by Dr. Genius, it will have a defensible meaning (though one, being genius-intended, that is likely beyond our ken).

Now imagine someone, Rene I’ll call him, trying to give evidence against the Dr. Genius hypothesis. Rene concedes that our seeing no meaning for a sentence isn’t reason to believe it has no Genius-intended meaning, since if it had such a meaning, we likely wouldn’t see it anyway. So he argues as follows: “If the sentence is by a Dr. Genius, then it has a defensible meaning. But our seeing no meaning to the sentence surely gives us darn good reason to suppose it has no defensible meaning. So it, too, gives us good reason to think it is not by Dr. Genius.”

Agnes: This seems like modus tollens. Yet something smells fishy with it, doesn’t it?

Bea: And not fresh fish either, I’d say. Notice that “This sentence has a defensible meaning” is less specific than “This sentence has a genius-­intended meaning.” The former is really a big disjunctive claim comprising various possibilities, one of which is the latter. To say “the sentence has a defensible meaning” is to say “it has defensible meaning written by a high school student OR it has a defensible meaning intended by an average undergraduate, OR by a bright philosophy major, OR an average graduate student,” and so on, til we get to what I’ll call “the last little disjunct,” namely: . . . “OR it has a defensible meaning intended by a Dr. Genius.

Agnes: That seems right. But how does it help?

Bea: Well, think of there being some threshold, such that if the evidence reduces the likelihood or credence-worthiness of a statement below that threshold, the statement is unreasonable to believe. Now Rene, while conceding that his evidence does not reduce the likelihood or credence­-worthiness of that “last little disjunct,” wants to bring the evidence to bear against the big disjunction. But the likelihood of the big disjunction is just the sum of the likelihoods of each disjunct (assuming they are, as here, exclusive). Hence, his evidence can’t make the disjunction unrea­sonable to believe unless, prior to the evidence, that last little disjunct was already unreasonable to believe. If that last little disjunct was worthy of belief before his noseeum evidence, then the big disjunction will be worthy of belief after the evidence, no matter how successful that evidence is in knocking out the other disjuncts.

Agnes: Would this be a similar example? Suppose I am trying to argue against Iris’s hypothesis that Bo killed her daughter. Lacking any direct evidence against this, I first point out that her hypothesis entails that one of her ex-boyfriends did it. I then amass evidence disconfirming this big disjunctive claim–evidence, that is, that Jack didn’t do it, that Tom didn’t do it, and so on for all her former boyfriends except Bo. This approach would be really fishy, wouldn’t it? For though my “evidence” might well reduce the likelihood of the big disjunctive claim, it couldn’t reduce it below the initial likelihood of Iris’s claim that Bo did it. So such evidence couldn’t provide reason to reject Iris’s claim. Is this similar?

Bea: Exactly. Athea’s argument tries to use her noseeum evidence as reason to reject the claim that the tragedy serves a God-sufficing good. But this too is just a big disjunction, one of those little disjuncts is that the evil serves some God-intended good. I argued that her evidence doesn’t reduce the likelihood of this crucial little disjunct. Athea conceded it doesn’t, but charged that this was a red herring. But it’s not, because what she concedes is intimately related to the issue of whether her evidence gives reason to reject the great big disjunctive claim that the evil serves a God-sufficing good. As in your analogy, Agnes, her evidence can’t make the big disjunction unworthy of belief unless, prior to her evidence, it was already unreasonable to believe the little disjunct. That’s why it was rel­evant to press the issue about God-intended purposes, although she didn’t couch her argument in terms of them. Shall we have some more decaf?

3.3 Coffee Pot #3

Agnes: Athea, you were about to say something about the Cartesian demon. Can we go to that now?

Athea: Okay. Bea’s analysis seems very insightful, though I want to think about it more. Perhaps I call adapt my demon point to it. Surely, taking everything into account, we are justified in believing there’s a table here. It’s not that our visual data makes it more likely that there is a real table than a demon table. But it does eliminate, or disconfirm, a lot of other possibilities–that there is a Sherman tank here, that there is nothing here, and so on. Now I think that when we get to the question of whether it is a real table or a demon table, the real table hypothesis wins out because it better explains why we have the visual, tactual, and other sensations that we do have.

Agnes: I think I understand, but I don’t trust arguments from evil demons. Could you give a more down-to-earth case?

Athea: Well, suppose we are at the zoo with a child, looking at what seems just like a zebra. The child imaginatively points out that donkeys might be cleverly painted to look just like zebras; she then wonders if some merry pranksters have done just that in the present case. Now if we had some good reason to think there were such donkey-painting pranksters, it might be hard to know what to believe about what we now see. But suppose we don’t have good reason to think this. In that case, given our visual data, reason requires us to believe there is a zebra before us. Our visual data does this, in effect, by making it far more likely that there is a zebra before us than that there is some osterich, hippopotamus, or other critter before us–it eliminates a lot of these possibilities, much as in Bea’s discussion of possible defensible meanings. To be sure, our data does not disconfirm the painted donkey possibility. But lacking reason to believe there are merry donkey-painting pranksters, this possibility doesn’t even get off the ground.

Now return to the God issue. Like our visual striped-critter data, our knowledge of suffering serving no discernible point rules out, or discon­firms, a number of possibilities: that this is a world without evil, for instance, or that it is one all of whose evil serves a point “near the surface,” or within our intellectual grasp. I guess Bea is right that it does not much disconfirm the claim that each of these evils serves a God-purposed point. But without any reason for thinking that God exists, that hypothesis is like the merry prankster possibility: it doesn’t get off the ground.

So I want to say that it really doesn’t matter whether evil dynamically or statically justifies us in believing that God doesn’t exist, any more than it matters whether our striped-critter data dynamically or statically justifies us in believing the animal in the zoo is not a cleverly painted donkey. Perhaps, as we converse with our imaginative child at the zoo, we are as justified in believing the no-painted-donkey hypothesis before having the striped visual data as we were after having it. Nevertheless, the data of seeing those black and white stripes serves an important evidential function in the conversation: it reminds us that we aren’t justified in believing that there are merry donkey-painting pranksters at work here. So also, in conversation with Descartes, our visual table data reminds us that we aren’t justified in believing the evil genius hypothesis. In reminding, it forces explicitness. Maybe the data of inscrutable suffering, I now want to say, serves a similar evidential function. Maybe all it does is force explicitness, reminding us that even before the data of inscrutable suffering, we are not justified in believing there is a God. But that is an important function.

[The women are quiet, reflecting on Athea’s words.]

Agnes: Why then, Athea, do you think that even before this data, we are justified in believing there is no God?

Athea: I still have doubts about whether a loving God would create so morally obscure a world as ours. But I think Bea is right about this: our seeing no sufficient good served by Carrie’s death does not count against the claim that there is a God who has made a universe with moral depth. But it now seems to me there is an analogue with creationism. Some creationists argue that the world was made five thousand years ago with signs of apparent great age–fossils and the like–already built into it. I guess the fossil record doesn’t count against this version of creationism­–it doesn’t “dynamically disconfirm” it. But we rightly reject this sort of creationism. So also, barring evidence for the view that God created a morally deep world, I think we must reject this view too. In general, shouldn’t we reject views that posit the existence of certain entities–­leprechauns, Loch Ness monsters, and the like–unless there is positive reason for their existence’?

There are deep issues here. To be justified, must all of our beliefs, be supported by evidence, or only some of them? If only some, which ones, and why? There are also questions about what makes one hypothesis a good explanation of certain phenomena-and about whether theism, to avoid irrationality, must provide a good explanation of certain phenomena.

Bea: If you are saying that belief in God requires adequate positive justification if it is to be justified, well, then I agree with you. Of course, I think it has the positive justification it requires–though I’m not, as a philosopher, as clear as I would like to be about what it is. There are, as you say, Athea, deep issues here.

But the issue we have been discussing is whether the inscrutability of evil counts against theism–whether, that is, it puts theistic belief in need of more positive justification than it would otherwise require. And I think the comparison with “old-earth creationism” points us back to the right issue here. The problem with old-earth creationism is that to protect itself from the fossil-data, a new addition has been tacked onto the basic crea­tionist claim. To be sure, the fossil data doesn’t lower the probability of this expanded version of creationism, but the expanded version has, by its expanded content, made itself less probable than it was. It has become that much more “top-heavy” relative to the evidence. And I take it that hypotheses are here like tables (real ones, not demon tables): the more top-heavy they are, the more support they require to stand.

But this really points us back to the issue Rowe had raised. Rowe, you remember, challenged Wykstra’s claim that theism itself makes it “entirely expectable” that our universe would have “moral depth.” The “moral depth” line, Rowe countered, is a significant “expansion” of theism, a kind of ad hoc addition. I think we are seeing that the issue concerning us tonight depends upon this. If Rowe is right in seeing the moral depth line as a significantly expanded version of theism, its immunity against noseeum disconfirmation is purchased at the price of making itself more improbable through added “top-heaviness.” I don’t think Rowe is right; but that does seem to be the real issue.

So we haven’t solved the problem; but I think we have narrowed it. For a while it seemed that even if theism entails (without “expansion”) that our universe would have moral depth, this wouldn’t help the theist, because of monstrous problems of circularity, demon tables, and the like. Those problems,” though interesting, now look more like shadow monsters.

Athea: Perhaps so. I too think we’ve made progress. But it’s way past midnight, dear friends, and I’m too tired to say anything with confidence just now. It’s time to go back to my place and turn in. We need sleep before we visit Iris in the hospital this afternoon.

Agnes: I just want to thank you both; I haven’t had a night like this since our college days. Remember how Iris loved our all-nighters? If only philosophy could help her through her pain now–instead of just giving me a break from it. And I still don’t know what I believe.

[Agnes looks out upon the street.]

But it seems a little less dark out there, doesn’t it? Do you know, I’ve never seen the sun come up. I’ve heard about “false dawn,” when sunrise seems near but is really still hours away. But I’ve never seen that either. Which is this, I wonder?


For the most part, Bea Leaver’s arguments come from S. Wykstra, while Athea Ist’s arguments come from B. Russell. Russell thanks Wykstra for laboring to improve the dialogical style. Wykstra thanks Russell for initiating and managing the project.

“The ‘Inductive’ Argument From Evil: A Dialogue” is copyright © 1988 by Philosophical Topics. All rights reserved.

The electronic version is copyright © 2000 by Internet Infidels, Inc. with the permission of Philosophical Topics, Bruce Russell, and Stephen Wykstra. All rights reserved.

all rights reserved