Tuesday, May 08, 2007

The Problem of Evil and the Burden of Proof

It seems a good deal of discussion on the problem of evil hinges on exactly where the burden of proof lies. If the burden lies with the theist to prove that the evils in the world can be squared with God's goodness, then the deficiencies that many of us find with the various theodicies on offer is a serious problem. However, the logical basis for saddling the theist with this kind of a burden seems to me to be hard to come by.

But I'm pretty much a skeptic about burden of proof claims. I think what has the burden of proof is whatever we don't believe, and it has that burden to the degree that we think that what we don't believe is unlikely to be true. If you want to argue that issue with me, maybe I'll set of my next entry so you can do that.

But, for the sake of argument, let us suppose that, all things being equal,

1) Theism is true.

has a burden of proof that

2) Atheism is true.

lacks, in virtue of the latter's superior ontological parsimony.

Even if we grant this, the problem of evil atheist is attempting to establish

3) The argument from evil shows that God does not exist.

In this context, the atheist is not content to use a lack of evidence/parsimony argument against the theist, the atheist is claiming that there is positive and decisive evidence against theism. Like it or not, it looks to me as if the atheist has picked up the burden of proof.

Andrea Weisberger has tried to argue that the theist does have the burden of proof on this issue. Graham Oppy, the agnostic reviewer for Internet Infidels, puts her claim as follows:

Oppy: In the Preface, Weisberger lays her cards on the table: she gives a clear and concise summary of the overall shape of her argument. She announces her belief that the success or failure of the theodicies which she investigates is pivotal for determining the success or failure of the argument from evil. In defence of this controversial claim, she appeals to considerations of burden of proof: it is the theist who is making an ‘extraordinary’ claim, and hence who faces the burden of squaring this claim with the ‘ordinary’ facts about the world. Moreover, she claims that to reject this burden of proof is to admit that the ‘extraordinary’ claim in question is ‘unfalsifiable’ (and hence, at least by imputation, not deserving of reasonable belief).

However, he thinks that is where Weisberger's argument fails:

Oppy: Perhaps the main fault which I find with the overall line which Weisberger takes lies in her appeals to the burden of proof. It seems to me that the right method here is to formulate the competing views--i.e. theistic and non-theistic theories of the world--and then to ask which one is best supported by the total available evidence. If theists can reasonably suppose that they have lots of evidence which supports the claim that God exists, then they may reasonably believe that there is a solution to ‘the problem of evil’, even if they do not know what that solution is. To insist, that theists have to provide a satisfactory theodicy or else abandon their theism, is to fail to pay proper regard to ‘the principle of total evidence’. Of course, this is not to say that non-theists cannot reasonably suppose that ‘the problem of evil’ helps to sway the weight of total evidence in favour of non-theism--indeed, I started out by claiming that this will very likely be the case--but that seems to me to be a very different issue.

I think Oppy's response is right on the money.

18 comments:

exapologist said...

I agree as well. If a theist is convinced by a set of arguments and evidence for theism, and it's sufficiently strong, then this evidence can make it rational to believe that God exists and has a sufficient reason for permitting suffering, even if they don't have the foggiest idea of what it might be.

Mark K. Sprengel said...

Maybe this is a bit off topic but I've recently run up against an atheist saying the burden of proof rests completely on the theist.

However, it seems to me that if one is rejecting an argument they're making an assertion about that argument and some burden of proof shifts to them in that case.

Anonymous said...

exapologist is quite right.

God has a sufficient reason to permit slavery , and the fact that nobody can think of what that might be is no reason to think that a theodicy could not be developed for slavery.

Anonymous said...

(NB: I am not the author of the previous comment.)

I must say that I detect more than a whiff of sophistry in Victor's claim that "what has the burden of proof is whatever we don't believe, and it has that burden to the degree that we think that what we don't believe is unlikely to be true."

Any belief can be cast in equivalent positive or negative formulations, and to shift the burden of proof onto the negative alone is both arbitrary and potentially absurd.

For example:
"I believe that theism is true." (No proof required.)
"I do not believe that atheism is true." (Burden of proof suddenly descends with an almighty crash.)

exapologist said...

Hi Anon,

The Christian need not abandon Christian theism flat-out even if they have sufficient moral justification to reject the stuff about slavery in the Bible.

Also, recall that my claim is a conditional one: if a theist has sufficiently strong evidence, then that can reduce the evidential force of certain worries.

I think this is a straightforward implication of the relationship between evidence and hypotheses in general. It's true in the sciences, it's true in crime scene forensics, and it's true here.

Anonymous said...

Vic, here's a fuller explanation from Weisberger about the burden of proof when she evaluated my debate with David Wood:

Another issue to be addressed is the burden of proof which, Wood claims, lies with the skeptic. But his position here appears to rely on a misunderstanding. The entire terms of the debate rests on a response to the god hypothesis, initially offered by the theistic view. The problem of evil, which is the focus of the debate, could not even arise unless a particular theistic worldview were presented beforehand.

This worldview, as Dr. Hatab noted in his introductory comments to the debate, is peculiarly western: god is assumed to be all powerful, (inclusive of all knowing), as well as wholly good. If we round out what these terms mean in their most profound sense, we should conclude that we are referencing a deity which is as powerful as logical possibility would permit, and so perfectly good that this being would be opposed to evil in every respect. So this god, no matter what other attributes might be claimed of it, would be powerful enough to eradicate evil (provided it was not logically impossible to do so) and motivated to do so by absolute goodness, which we can suppose is the opposite of evil.

If we posit the existence of an omnipotent (and omniscient), and omnibenevolent deity, then one might wonder why there is such an abundance of suffering or evil in the world. It is only if we posit the existence of such a god that evil becomes a "problem." So we see that the questioning of the existence of god, or the plausibility of the god hypothesis, only occurs in response to the god hypothesis. As a result, the burden is on the proponent of the hypothesis or the presenter of the initial claim.

An analogy would be if someone were to claim that invisible, green gremlins power all microchips. Confronted with this hypothesis, one might ask how this is so, how it is known that these gremlins are green if they are invisible, and many similar questions. It is simply not convincing for the proponent of the invisible, green gremlin hypothesis to then claim, 'Well, since you question the gremlins' greenness, it is up to you to prove that they are not green!' This does little to persuade anyone of the viability or plausibility of the gremlin hypothesis. Similarly, anyone making claims about the existence of extraordinary phenomena, such as invisible, green gremlins, the burden of proof lies with the claimant. And the claim about the existence of a wholly good, all powerful being, in the face of such abundant and excruciating suffering in the world, appears to be an extraordinary claim! It is the proposer of the god hypothesis, no matter what the flavor (classical or personalist), who must bear the burden of making sense of the claim that an all good, all powerful being -- one who is powerful enough to eradicate at least some of the tremendous suffering that exists, and one who is opposed to such suffering by its very essence -- exists.

I'd appreciate your thoughts on this.

David Wood said...

I'd also be interested to know what people think about this issue.

Reply to Weisberger on Burden of Proof

Reply to Weisberger on Burden of Proof II

Darek Barefoot said...

Consider the claim that it is logically necessary for God to allow some suffering to occur temporarily to bring about a world in which suffering does not occur. How much actual suffering would be inconsistent with this claim? One tenth the suffering that is now occurring? One millionth? If only one millionth the suffering occurred, is it possible that it would seem to us that too much suffering was still occurring to admit the claim? If only one instance of suffering in the world were occurring, and the instance were my own agonizing toothache, it would seem like a lot of suffering to me. I just don't see what means we have available for knowing how much suffering rules the claim out.

Anonymous said...

Uh Oh. David Wood is here. Oops. Don't want to do that again. ;-)

But I have alerted Dr. Weisberger to this discussion. I hope she comments.

Blue Devil Knight said...

I don't think this is all that complicated, even though having the burden of proof is not a crisp category.

People who say outlandish things, given the priors of what reasonable and educated people believe, have a burden of proof. There is no a priori way to say, then, whether position X or Y carries the burden.

Sometimes it is obvious who has the burden (e.g., someone wanting to say Earth is flat or that Zeus exists). Sometimes it is obvious someone doesn't have a burden. Sometimes it depends who you ask (a group of Christians at Jesus Camp will clearly put the burden on the poor little boy who stands up and questions whether there is a God; while a group of scientists at a conference on the retina will think someone who stands up and says 'God makes retinal ganglion cells respond to stimuli' will think he has the burden of proof (at best that is what they will think)).

I think it can be handled in a Bayesian fashion, and this is how it plays out amongst philosophers. Basically, once a philosopher makes a case for something that everyone thinks works well, someone gainsaying his claim will have the burden. E.g., since Twin Earth people who think that contents are fixed internally have a big burden.

We can't answer the question who has the burden without knowing the priors of all the parties involved. With the God question, it becomes clear that the priors are so different on different sides of the fence that they will tend to disagree about who has the burden.

Sturgeon's Lawyer said...

H'mmm. First, I agree; the Problem of Evil is a separate matter from the Burden of Proof.

However, when the PoE is phrased in the classic manner of the riddle of Epicurus, things become a bit murkier.

Is God willing to prevent evil, but not able? Then he is not omnipotent.
Is he able, but not willing? Then he is malevolent.
Is he both able and willing? Then whence cometh evil?
Is he neither able nor willing? Then why call him God?


This takes two hypotheses -- there is a being called God, and there is this thing we call Evil -- and maps out the four logical possibilities that follow from it. If we, as theists and Christians, assert that there is a God, Who is omnipotent and benevolent, then this mapping creates a burden of proof, not of God's existence, but that the existence of the kind of God we assert is logically consistent with the existence of the thing called Evil.

Theodicy is, of course, the response to this burden of proof; and it is disturbing that no theodicy to date has ever been really satisfactory (nor do I have a better one to offer)

exapologist said...

I think some of this discussion is diverging from Oppy's (and, I take it, Talbott's) point, which was a conditional. Some of the discussion is assuming that the antecedent is being defended here (viz., that theism, construed as a large-scale hypothesis, is made significantly more probable than not with respect to the relevant evidence). Of course, Talbott and other theists think the antecedent is true, but that's not the point being raised here. And on this point, I think Victor is correct.

So suppose the posterior probability of (classical) theism (minus the data of suffering) is .8 or .9, or thereabouts. Then the theist may be justified in thinking that God has a reason for permitting evil, even if they don't know what it may be). The data of suffering may lower the probability of theism when added in, but it might not lower it below 1/2.

Things may be different, of course, if the posterior probability of classical theism (minus suffering) is much closer to 1/2 to begin with, but I don't think that's the sort of scenario Oppy (or Victor) has in mind here.

exapologist said...

Sorry, "...Oppy's (and, I take it, *Victor's*) point..."

Anonymous said...

I understand exapologist. But I have three questions:

1) How does one go about rationally assessing the posterior probability of classical theism (minus suffering) in the first place, when there are no mutually agreed upon tests to decide between religious (and probably) non-religious beliefs? I argue that our beliefs are overwhelmingly adopted because of when and where we were born, and hence cause for strong initial doubt.

2) When it comes to theism as a whole (including the POE) isn't Weisberger correct to say, as she did, that such a view as a whole, "appears to be an extraordinary claim"?

3) Just because a believer may conclude theism has an initial probability of, say, .8 or .9 (minus the POE), it doesn't mean in a debate with someone who disagrees that she can throw the burden of proof upon her opponent, correct? [In a debate format, the language of the proposition to be resolved will express who has the burden of proof, and Weisberger was partially responding to that issue.]

Until I read what Weisberger wrote I had always maintained that each side has its own burden of proof to meet. Now I lean to her side, but I don't think it solves anything to debate who has the burden of proof, since we must still offer our arguments pro and con. This issue is just another one of many though, and worth the effort.

Cheers.

exapologist said...

Hi John!

I think I see what you're getting at. in debate contexts, sometimes the following sorts of claims get conflated:

1. *You're irrational* to (dis)believe that P.

1'. *I'm not irrational* to (dis)believe that P.

2. *Anyone* who (dis) believes that P is irrational.

As far as I can make out, Oppy and Reppert are talking about (1')-type claims. If so, then it seems that the burden's a bit lighter than it would be if the theist (or atheist) were asserting a (1)-type claim (at the very least, it's a lighter burden than exists for a (2)-type claim). When it comes to claims of the latter sort, I tend to agree with you about the points you mention.

Hope all's well with you and yours!

EA

Anonymous said...

But the real question is, who owns the burden of proof to show who has the burden of proof?

exapologist said...

:)

Jason Pratt said...

Anon's reply was pretty good. {g!}

I've always understood the burden of proof to be on the one who is trying to score the point, whichever side that is, whatever the circumstance is. Naturally, in a highly favorable environment, the point can be perceived to be 'established' without a lot of 'work' being done. Even so, the principle remains the same.

Consequently, I just consider myself to have the 'burden of proof' in all cases. Similarly, if I want the judge to consider me to have the right-of-way on the strip, so that a score will be counted if I land on target, then that's my responsibility. Not the judge's. Not the other fencer.

After all, the 'problem' with burden of proof, only happens when we're really asking, "when can it be the other guy's problem and not mine?"

Never. If I'm going to be fair to the opponent, it's _always_ _my_ problem. The end.

Incidentally, this also answer's Anon's perceptive quip. {g} {bowing in admiration in his or her direction!}

JRP