Friday, March 07, 2008

The Logical Problem(s) of Evil

Many theists seem to treat the 'logical' problem of evil as refutable simply by establishing that God's existence is compatible with (some possible instance of) evil. They then infer that the only real challenge is from the 'evidential' problem of evil, which even if it works merely shows there is some (pro tanto, defeasible) evidence against God's existence. So, nothing too threatening.

But this is poor reasoning. The mere fact that God's existence is compatible with some possible evils does not establish that we have only probabilistic evidence against his existence. For, obviously enough, it does not entail that God's existence is compatible with the particular evils we actually observe. Indeed, I think there's a very strong case to be made that God's existence is logically incompatible with the particular evils we observe.

Since my conclusion is that God's existence is impossible, not merely improbable, it would seem misleading to call this an 'evidential' argument from evil. It is properly considered a logical argument from evil. But it certainly isn't refuted by yelping, "free will! free will!" So it's a mistake for theists to claim that the logical argument from evil has been refuted. (At best, all they've refuted is the most simple, straw-man version of the argument.)


  1. Exactly!

    Btw, have you read Lewis' brickbat on evil?

  2. Hmm, not sure. (I vaguely recall some letters he exchanged with Adams.) Is it online?

  3. "Indeed, I think there's a very strong case to be made that God's existence is logically incompatible with the particular evils we observe."

    Or at any rate incompatible with the existence of a _benevolent_ God. But perhaps your arguments can be considered probabilistic evidence (albeit incomplete) _for_ the existence of an evil God? ;)

  4. (At best, all they've refuted is the most simple, straw-man version of the argument.)

    It can't be a straw-man version of the argument because you can find real versions of the argument that are refuted by the move. (It's unfortunately a feature of much theist-atheist discussion that however crude an argument may be, you can usually find someone on either side who affirms something very much like it.)

    I'm inclined to agree with your basic point. But if God's existence is compatible with some possible instance of evil, then the argument moves from the consideration of evil as such -- because if the response is sound, there is nothing about evil as such that is incompatible with God's existence. Rather, what has happened is that one has shown that either 'the logical problem of evil' is really just an evidential problem of evil, or that the phrase is in fact misleading: it is not evil as such, but some particular feature, F, whatever it may be, that attaches to some possible kinds of evil. And, of course, what can be said about that depends on what F is -- and you'll be hard-pressed to find any statements by atheists as to what F is that are both clear enough to work with and generally accepted enough to be well-known.

  5. I find it a little confusing to talk about evil refuting the existence of god because traditionally the word god was not directly associated with benevolence. In fact I don't think it is even now for most people in the world.

  6. That's easy to fix: replace every instance of 'god' with 'omniscient, omnipotent, and morally perfect being'.

  7. It doesn't seem to me that any particular instance of evil could be considered a defeater for god's existence (assuming, for the moment, we're talking about the Christian god), given the conception of god -- and of his purposes -- most Christian theologians subscribe to.

    Most would first point out the fact that (according to the their dogma) god's purpose isn't to make us "happy" and "comfortable." Quite often, removing this erroneous premise alone does quite a bit of work when discussing the problem of evil.

    Second, most point to human experience. When we think about our lives, or about the lives of others that we most admire, they're never the happy, comfortable lives some people suppose a wise god would want for us, but lives full of challenges and, it must be said, much suffering. If even we, on reflection, can see the fact that full, admirable human lives necessarily include suffering (think of "Brave New World"), is it honestly difficult to conceive of a wise god seeing things this way as well?

    A third point they often raise is related to the second. God is often compared to a father figure, and it is incontrovertible that wise, moral fathers are willing to let their children suffer in certain circumstances, even when it's within their power to prevent such suffering. It is also important to recognize the fact that, at the time, the child often doesn't understand her father's purposes, and may indeed wish that her father would do whatever he could to help her. If this is so, then it's at least prima facie plausible that god could be morally perfect (however defined) while allowing human beings to suffer in ways that they would prefer to avoid, for purposes they may not understand.

    This third point sheds some light on another error many are guilty of when thinking about suffering. They think about god, and about either evil in general or about particular instances of evil in particular, but they often forget to think of these things in relation to Christian teachings about human beings.

    I would like to add that I am not a Christian, I'm an unbelieving agnostic (which I suppose makes me a de facto atheist). But I think that many of the arguments presented against various theistic beliefs are often not as forceful as they at first appear.

    Finally, and off topic, I'd like to say that this is my favorite blog! I hope that you continue blogging once you become a professor, Richard!

  8. "replace every instance of 'god' with 'omniscient, omnipotent, and morally perfect being'."

    I don't think that'd work either. You can read the Book of Job as saying God is not omniscient, omnipotent and morally perfect. God is just God, and the message there is how to cope with a God that's nothing you thought He'd be. It is a "reality principle" of sorts.

  9. Jared, I think Richard's point in his reply to genius was precisely that he is only interested in refuting (indeed, logically refuting) the possibility of a god who is "omniscient, omnipotent, and morally perfect" (three characteristics that are sometimes, but not always (and according to genius, not even often), ascribed to god); he is not trying to refute every possible conception of god that can be found throughout history, such as, e.g. the conception of god that you find in the Book of Job.

    Eric, I agree. This is quickly becoming one of my favorite web sites.

  10. Hey Richard,

    It's not available online so far as I know, but there's an excerpt in Harper's and a longer version in the collection _Philosophers without Gods_.

    Quick note to Eric,
    I'm not quite sure to make of the claim that there are no particular instances of evil that count as defeaters for God's existence. Defeaters are typically bits of evidence that undermine or override other evidence one has for a belief that can potentially lead to a loss in justification. Arguments from evil are not intended as arguments to show that Christians are irrational or would be irrational if only they considered our arguments. The argument from evil is an argument for the non-existence of God. The argument can be conclusive even if not rationally recognizable as such by all people.

    My view, fwiw, is that the Christian teaching that you describe may well put someone into a position where they cannot appreciate the rational force of the argument from evil because it has essentially given then a warped view of morality that only years of therapy could undo. It leads them to say insane things like (a) there aren't conclusive reasons to fetch unwatched drowning babies from backyard pools or (b) that there are good reasons not to alleviate the suffering of a woman who has just been gang raped and had her limbs hacked off by machetes. When we see what these people are committed to believing by virtue of their Christian teachings (i.e., that there are overriding reasons for non-intervention in such cases), I could care less personally that there are no 'defeaters' that I could offer that could address them where they are given their perspective and bring them to the light. Some people you just cannot reason with. I'd include scientologists and young earth creationists on that list and those who are prepared to say things like (a) and (b) above in response to the logical argument from evil.

  11. This is something I've also recently been thinking about. However, as the most common theistic objection to this sort of argument is the "mystery" reply, that God has some unknown reason for allowing particular evils, I think that the case can only be made that theism is probably incompatible with the particular instances of evil.

    In this case, how unlikely are we to judge theism? Should we believe that theism is impossible because it probably is, or should we merely hold it as wildly improbable?

  12. As a comment to Derrida's comment, one may ask what exactly probability means in a context like this. What does it mean to say that god "probably" doesn't exist, or that it's "wildly improbable" that he does? (And what is the difference in degree in these alternatives supposed to mean?)

    Does it mean that if we had 1000 other worlds exactly like our own, god _would_ exist in only 5 of them? Clearly that's an absurd assertion. It's wrong to try to assign a numerical probability to the existence of god, or even to suggest that such a probability can be assigned in theory. We obviously have in mind a vaguer notion of "probably" than a numerical/mathematical one -- but what is exactly does this notion consist in?

  13. bv: When people talk about "probability", they can mean one of two things:

    - Objective probability; how likely an event is to occur given the laws that act upon it.

    - And epistemic probability; how likely an event is given what we know about it.

    So, a coin coming up heads, given good weather, may have an objective probability of 50%. But, the epsitemic probability of the event given that we have witnessed the coin coming up heads will be significantly higher.

    In my previous comment, I used the term "probability" in the sense of "epistemic probability", how likely something is given what we know.

    In effect, I am asking how likely the existence of God is, given the evidence presented by Richard in the opening post.

    Your question, (Does it mean that if we had 1000 other worlds exactly like our own, god _would_ exist in only 5 of them?), is asking about objective probability, which I would agree in the case of God's existence is incalculable to say the least.

    In regards to your other question, how improbable is wildly improbable, I was in effect asking what Richard's argument from evil does to the epistemic probability of God's existence. Because we now have good grounds, though not conclusive grounds, to think that God's existence is incompatible with the facts, how is God's epistemic probability affected?

    If God's (epistemic) probability still isn't naught, then is it an infinitesimal leap away from naught? Or is it in a certain arbitrary place between naught and 0.5? Can such calculations be made?

    My question was in effect a vehicle for further inquiry into the question of epistemic weight with regards to Richard's variation of the logical problem of evil.

    Hope that helps :-)

  14. Thanks for the reply, Derrida. Please bear with my extensive reply.

    Until recently, most mathematicians and physicists agreed that what you here call "epistemic probability" is the only kind of probability there is. Probabilistic reasoning was seen as a heuristic method that provided an efficient way of approximately predicting future events based on an incomplete knowledge of the relevant data (i.e. past and present events, as well as physical laws). The idea that there might exist "real" chance out in the world was not something that a lot of people took seriously.

    This idea has since come under some pressure. Bell's Theorem showed that every local theory of the world on a small scale (i.e. a theory that assumes, roughly, that you cannot have information travel faster than light) must necessarily be statistical, and this has led many people to conclude that nature itself must be statistical, at least on a small scale.

    Nevertheless, we have no reason to suspect that determinism does not rule on a macroscopic scale, and it does seem that most probabilities we encounter outside of theoretical science is of the epistemic kind. In the case of a coin toss, we say that the coin has a 50-50 chance of coming up heads (either based on our past experience with coin tossing, or based on the symmetry of the coin and assumptions about the regularity of physical laws), but presumably, if we had more information about the specific coin toss, we could find a better probabilistic estimate, or perhaps even ascertain how the coin would land.

    Now, I agree that the way we use (epistemic) probability to assess the likelihood of god's existence is different from the way we use it in the case of the coin toss, but the meaning of what we are doing still escapes me, and I'm afraid your explanation doesn't help me much. I know that you were trying to assess how God's epistemic probability affected is affected by Richard's arguments, but I have no idea what this might mean.

    You seem to still want to even talk about numerical probabilities, but to me this seems tied to the kind of mathematical analysis we did for the coin toss. The analog of that kind of reasoning for the god case _would_ be something like "If we had 1000 other worlds exactly like our own, god would exist in only 5 of them". In this case, the probability has a _meaning_: it is a ration. But when you say this, I have no idea what you mean:

    > "If God's (epistemic) probability still isn't naught, then is it an infinitesimal leap away from naught? Or is it in a certain arbitrary place between naught and 0.5?"

    What would 0.5 mean here? (or any other number?) Divorced from the kind mathematical analysis we did for the coin toss, it would certainly no longer be a _ratio_. Indeed, it seems to me it would no longer signify anything but an ordinal on an ordinal scale: that is to say, you might as well ask, 'given Richard's arguments, did god's probability go from "A" to "B", or did it go to "C"' (if we give the alphabet the obvious ordering). That is to say, the question would no longer be very interesting, or even meaningful.

  15. Perhaps we shouldn't talk about probability as such; in my haste to put across my views, I may have assumed too much.

    Let's put it instead in terms of believability. How believable is the proposition that God exists. If I were to take God's existence as believable as the truths of mathematics, which seem to me incorrigible, then I would put God's existence at the top of a believability scale. Let's call the top of this scale "one", and the bottom of the scale (On which reside such concepts as married bachelors), "zero".

    To say that God is at "0.5", or halfway down the scale is to say that we should neither reject it or accept it, as our reasons for doing either are equally balanced.

    Now, assume for the moment that believability should correspond to the truth of a proposition, such that if we think a proposition to be true, we should correspondingly believe it to be true to the extent which we have reason to do so. If a piece of evidence makes something half as believable as it would otherwise be, then it correspondingly moves down the scale.

    This seems not to entail your suggestion, specifically that we should believe God to "occur" in five out of a thousand worlds, because the believability account uses a completely internal measure of evidence and argumentation.

    How does a piece of evidence affect believability? Using a simple Bayesian account of evidential support, evidence E should make Hypothesis H more believable if we would expect E given H, and not expect E given ~H.

    Now, because we would not expect evidence of unjustifiable evils given God's existence, and because unjustifiable evils are in fact inconsistent with Theism, this seems to mean that the proposition "God's existence is contradicted by the facts of the world" is eminently believable. But what does this do to the proposition "God exists"? How believable is it now? Is it near the lower end of the scale, or smack on zero?

    On a side note, I'd be greatly interested to learn about your own views on the above argument from evil, and on probability in general. If all probability measurements are based on meaningless propositions, ("What does it really mean to say that something is "0.5" likely to happen?!"), then does this mean that the concept of probability is doomed to failure, and that as such nothing can be claimed to be likely or unlikely?

  16. Derrida - I agree that we can sensibly talk about subjective credence (or rational 'degree of belief'), as a purely epistemic term that doesn't have the implications BV worries about. In particular, one may accept that God's existence is logically incompatible with our actual evidence, without being epistemically certain of this metaphysical judgment. So non-zero credence may be justified. It's a separate issue.

    Others - BV is right that I'm only interested in refuting the existence of a truly perfect being. Of course there could exist a being that morally deluded cultists believed was perfect. That's quite consistent with the evil we observe. (I also agree with everything Clayton said.)

    Eric and BV - thanks for the kind words and encouragement!

  17. Richard,

    Sorry for dragging this up (I found it via your review of the year), but I can't see how your version of the logical Problem of Evil can't be refuted by Plantinga's version of the free-will defence. Assuming Plantinga's FWD is successful against say, Mackie's logical PoE, how is it unsuccessful against yours?

  18. I take the FWD to claim that omnimax God's existence is compatible with some 'moral evils' (which it is claimed must be tolerated as a necessary condition for individuals' free will -- a claim I've refuted elsewhere). But even if we grant that, it doesn't necessarily justify all evils. God's existence might still be logically incompatible with so-called 'natural evils', for example.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)