Performing Contradictions and the Problem of Evil
September 19, 2006 — 10:49

Author: Jeremy Pierce  Category: Concept of God Existence of God Problem of Evil  Tags: , , ,   Comments: 25

I’ve been teaching an introductory philosophy course this semester with a new text for my God unit, Thinking About God by Greg Ganssle. It’s designed to be usable for high school or introductory college/university courses, and it’s just about the lowest level of detail that I would want to use for this course. I’m supplementing it some with other readings also, but it’s nice to spend a lot of time just in one book after using lots of scattered readings in past versions of the course.
One thing that I found really interesting was in the section on the logical problem of evil. The logical problem of evil presents three traditional attributes of God (omnipotence, omniscience, and perfect goodness) and then seeks to derive a contradiction if you admit to the existence of evil (which pretty much all traditional theists will do, and thus it’s a problem even if the person presenting the problem doesn’t happen to believe in evil, because the theist does, and it’s supposed to be a contradiction for theism). Now it so happens that hardly any philosopher today accepts the logical problem of evil as a good argument, for several reasons, but in the process of explaining why Ganssle hits on an interesting issue that I hadn’t thought of before. One way some people have resisted theists’ attempts to respond to the problem of evil might actually help the theist in surprising ways.

Responses to the logical problem of evil can involve explaining why a perfectly good, omniscient, omnipotent being would allow evil. For instance, free will is given as something important enough that God would want it, even if it means a fair amount of evil would be allowed. One kind of response to that (not taken seriously by most philosophers) is that if God is omnipotent then God should be able to give free will and also guarantee that people will freely do no evil. The standard response is that libertarian free will is incompatible with God guaranteeing what people will do, and God can’t perform a contradiction, since contradictions are impossible. So God can’t both give free will and guarantee what people will do, since guaranteeing what people will do violates free will. That would amount to performing a contradiction.
But isn’t God omnipotent? Doesn’t that mean God can do anything? The traditional answer is no. Rene Descartes is an extremely rare exception to the overwhelming consensus among theists that God cannot perform contradictions, because there is no such thing to be performed. Maybe you can define something called superomnipotence and then say that superomnipotent beings would be able to grant free will and then guarantee what people will do, but that’s not the sort of thing theists hold to, because God is merely omnipotent. In fact, nothing could be superomnipotent anyway, and claiming that God is superomnipotent would already be claiming and impossibility.
What I found really interesting in Ganssle’s discussion of this is that he thinks the superomnipotence objection to the problem of evil actually counts in the theist’s behavior. What if the theist were to concede that God is superomnipotent? You might then think that God doesn’t have a good reason for allowing evil anymore, since God could perform the contradiction to guarantee people’s choices while maintaining their libertarian free will. But not so fast. Does God need a reason to allow evil if God is superomnipotent? Superomnipotence means God can perform contradictions. That means God can allow evil even if evil contradicts God’s goodness, omnipotence, and omniscience. If you’re going to allow contradictions with a superomnipotent being, then why is it at all problematic that God and evil contradict each other? This objection seems to fall flat if you allow God to be superomnipotent.
I had never considered that point before, and while I think Ganssle is right I do want to say one further thing. Once you allow contradictions, all logic goes out the window. Contradictions logically entail that every statement is true. So it’s not going to be surprising if it turns out that the theist is not threatened anymore by the problem of evil, since every statement turns out to be true, including that one. But it’s also going to be true that God is unjust and evil, since that’s also going to be a statement, and every statement follows from contradictions. Once you allow contradictions you can prove that God doesn’t exist just as much as you can prove that God is vindicated in allowing evil. In the end, I don’t know if Ganssle’s point establishes very much, since any statement follows as true once you allow contradictions. I do think it was an interesting observation, however. There’s a reason hardly any philosopher has endorsed superomnipotence as a plausible interpretation of omnipotence. It’s completely ridiculous to suppose that theists have ever meant that God can do something that makes God both exist and not exist, so omnipotence could never have meant that.
[cross-posted at Parableman and Philosophy et cetera]

  • Kevin Timpe

    Yes, Harry Frankfurt says something similar in response to the ‘stone’ objection to omnipotence:
    If God is supposed capable of performing one task whose description is self-contradictory—that of creating the problematic stone in the first place—why should He not be supposed capable of performing another—that of lifting the stone? After all, is there any greater trick in performing two logically impossible tasks than there is in performing one? If an omnipotent being can do what is logically impossible, then he can not only create situations which he cannot handle but also, since he is not bound by the limits of consistency, he can handle situations which he cannot handle (Frankfurt, “The Logic of Omnipotence,” 63).
    If I remember correctly, he then also suggests that the same line of response is available for the problem of evil. Paraphrasing here, “The existence of evil is logically incompatible with existence of the an omnipotent, omniscient, omnibenevolent God? No problem. God can bring about both.”

    September 19, 2006 — 11:41
  • Kevin, that’s exactly how John Hawthorne describes Descartes’ view. God would be capable of making such a stone and of lifting it. I didn’t know that Frankfurt had applied this to the problem of evil. Is that a paper or a chapter in a book?

    September 19, 2006 — 17:02
  • Kevin Timpe

    Again, I could be wrong that he makes this connection. But I think he does. Here is the article: Frankfurt, Harry. “The Logic of Omnipotence” Philosophical Review (1964).

    September 19, 2006 — 17:23
  • I heard van Inwagen make a similar point in a lecture. But if we take seriously the idea that God can do contradictory things, we need to take dialetheism seriously – the view that there are true contradictions: see Graham Priest: In Contradiction, Is Truth A Liar. I don’t agree with his view that contradictions can be true, but if I did, I’d also agree that we need a paraconsistent logic – in which a contradiction does not imply any arbitrary statement. I’d also want to say that if it is true that there is evil and there is no evil, still that includes the truth that there is evil, and if God exists there shouldn’t be.

    September 20, 2006 — 9:59
  • Dear Jeremy – this is an exciting topic, at least I find it exciting. So exciting, in fact, that I wrote my above comment in haste, just before dashing off to teach a class. I’d like to take the chance to express myself a little more clearly.
    Graham Priest distinguishes trivialism – the view that everything is true (and therefore, everything is false, because one thing that is true is ‘everything is false’) from dialetheism – the view that some, but not all contradictions are true. In classical logic, dialetheism implies trivialism, because from a contradiction everything follows. Priest however does not accept this law of classical logic, and proposes that we accept some form of paraconsistent logic, in which the rule of ex contradictione quodlibet is abandoned.
    From a theological point of view, it is tempting to find here a way of explaining doctrines such as the Trinity and the Incarnation. Of course, I know that there is a long literature devoted to finding consistent explanations of these doctrines. In any case, as a philosopher, I am not convinced by Priest’s arguments that contradictions can be true. Still, it seems to me that if we are seriously considering superomnipotence, we should do so from the perspective of dialetheism and paraconsistent logic.
    So, as I understand Priest’s position, the situation is as follows. If I have a true contradiction, it is also, of course a false contradiction: a true contradiction is a proposition that is both true and false. We can deduce from it all the things that follow from its being true, and all the things that follow from its being false. We do not, however, have any license to deduce any arbitrary statement from the contradiction.
    The logical argument from evil rests on the claim “If there is evil, then God does not exist.” But we argue, perhaps God, in his superomnipotence has brought it about that “There is evil in the universe and there is no evil at all in the universe.”
    “Well” says the atheologian, “I’ll take seriously the possibility that there is evil in the universe and there is no evil at all in the universe, (although I’d like to see some evidence for the latter claim). Still”, says the anti-theologian, “from “There is evil in the universe” we can deduce “God does not exist,” and, after all,”argues the anti-theologian,” “There is no evil in the universe at all” does not imply “God exists,” at best it implies” (argues the atheologian) “”There is an epistemic possibility of God’s existence unless one discovers some reason why God cannot exist.” ”
    Incidentally, there is an article here:
    about the possibility of applying paraconsistent logic to Christian doctrine: the author is against doing so. I have to be frank and say that in my opinion, Quodlibet sometimes publishes articles that are simply below the standard one expects from a peer-reviewed academic journal, but I think this article is well worth reading.

    September 20, 2006 — 11:53
  • The phrase “performing a contradiction” also raises an interesting problem. Presumably what is meant is that God cannot perform an action whose description is inconsistent. But inconsistent in what logic? Suppose we choose classical logics. In that case God cannot create an object that is neither blue nor not blue, since it violates the excluded middle theorem that (x)(Bx v ~Bx). But that’s absurd, since there are objects that are neither blue nor not blue–they are for instance indefinitely blue. So there is this reason as well to move away from classical logic as a criterion for omnipotence. I guess an epistemicist could insist that every object is either blue or not blue. But the cost is that, with regard to some objects, not even God knows which color the object is. That unknowable truth might be a problem for divine omniscience.

    September 20, 2006 — 13:09
  • How is epistemicism supposed to imply that God doesn’t know what color the object is? It certainly doesn’t on Williamson’s version of epistemicism. I don’t know any other versions very well, but he’s pretty clear in his book that the inability to know has to do with human limitations that wouldn’t be true of an omniscient being. The facts that serve as the truthmaker of whether it is blue are facts about human language, human perception, and so on that are so complex as to be impossible for humans to figure out where the lines between the blue and the not blue would be. But that doesn’t mean God wouldn’t be able to do it.

    September 20, 2006 — 21:02
  • Steven Carr

    Can God create beings like the angels Gabriel and Michael that He knows in advance will never use their free will to choose evil?

    September 21, 2006 — 3:35
  • Just a reference:
    Earl Conee’s “The Possibility of Power beyond Possibility,” PHILOSOPHICAL PERSPECTIVES, 1991 explores several issues involved in what you’re here calling “superomnipotence.” Section VI is on the problem of evil. I *think* I remember that Conee thinks that this strong view of omnipotence makes the problem of evil harder for the theist.

    September 21, 2006 — 8:23
  • Steven, my personal answer is of course yes, since I’m a compatibilist, but I don’t think you’re asking about what I personally think. Can create libertarianly free angels who God knows will never do evil? That depends on whether God has middle knowledge, doesn’t it? If they are libertarianly free, and God does have middle knowledge, then can’t God just create them and place them in circumstances in which they won’t do any evil? Unless there are none, which goes back to the transworld depravity issues, it does seem to be a problem for Molinists who use the free will defense against the problem of evil.
    It’s an interesting question, because it seems as if what might help with maintaining traditional views on foreknowledge might hurt a little in dealing with the problem of evil. But we already know that that’s true with compatibilism, which makes the foreknowledge problem a piece of cake while having to answer the problem of evil in a way that doesn’t appeal to free will. So it’s not as if Molinists are here going to be in new territory if they can’t answer this question the way they might want to. It doesn’t leave them without any resources. It just means the main point of being a libertarian seems not to help after all, and they might as well have been compatibilists (at least in terms of the problem of evil). But again this all depends on what you think of transworld depravity and related issues (which I don’t want to get into again here, since I know what you think of those, and I know what the Molinists here think of those).

    September 21, 2006 — 8:52
  • I guess that’s supposed to be a reply to me. But, once again, I have no idea how it’s relevant to what I said. I said nothing even even close to “epistemicism implies that God does not know what color an object is.” What I said is that it *might* be a problem for divine omniscience. It clearly matters crucially here how you read the modal claim that some propositions are unknow-able. Could intelligent aliens know? Could divine beings know? I’m not certain. And of course I never said I was certain.

    September 21, 2006 — 9:25
  • Mike, I don’t want to get into a fight about what you meant, but what you wrote looked to me (and still looks to me) as if it’s saying that epistemicism might get around the problem but with a further cost — that even God wouldn’t know which color the object is:

    I guess an epistemicist could insist that every object is either blue or not blue. But the cost is that, with regard to some objects, not even God knows which color the object is.

    All I was saying is that epistemicism (in the form one of its most prominent proponents puts it) does not have that consequence. I can’t see at all how you might think that’s irrelevant to what you said, since it seemed to me (and still seems to me) that it deflects exactly the worry you give for using epistemicism in the way you suggest.

    September 21, 2006 — 16:26
  • Fair enough. Williamson’s epistemicism does not come down either way on what beings with greater cognitve capacities could know, and I didn’t mean to do so, either. Williamson is explicit about it in section 7.7 Decidable Cases, p. 212, in his (1996). For some reason I think Sorensen (maybe in his ‘Unknowable Obligations’?) says something stronger, but I’ll have to check.

    September 21, 2006 — 16:42
  • Yes, Sorensen, ‘Unknowable Obligations’ Utilitas ’95 argues that some obligations are ‘logically impossible’ to know. They are necessarily indeterminate. That’s inconsistent with at least some views on omniscience. Williamson too urges that a strong version of epistemicism would make omniscience impossible by making ignorance *essential* to indeterminate propositions. See section 6, (1996), Omniscient Speakers. In any case, whether epistemicism entails non-omniscience depends on the form of epistemicism in play.

    September 22, 2006 — 11:35
  • Hmm. I haven’t had a chance to look at Williamson again, but I remember discussing this issue with John Hawthorne, and he seemed to think Williamson’s view didn’t require saying that. Maybe he was just saying that epistemicism need not say that and was offering it as a correction to Williamson, but I remember it as him simply giving Williamson’s view.

    September 22, 2006 — 15:07
  • I have had a chance and (for convenience) list the references above. Williamson’s view has a few sides to it. But he does say that *he does not know* what beings with greater cognitive ability would know regarding indefinite propositions. There are two possible positions, as far as I can see. Either epistemicism entails that beings with greater cognitive ability do not know the actual truth value of indefinite propositions or it does not entail that. If it turns out that beings with greater cog. ability do not know such truth-values, then epistemicism does have that entailment. Otherwise not. But no one (save Sorensen on unknowable obligations) seems to know.

    September 22, 2006 — 16:06
  • Christian

    Jeremy & Mike,
    I’m a bit confused about the debate and how epistemicism is involved in it. Isn’t epistemicism, primarily, the view that there are sharp cut-offs as to whether certain vague predicates apply to a given object “and” that we humans cannot know, in principle, what those cut-offs are? This would then be silent on the question as to whether there is any predicate F such that “possibly, I(Fx)” where ‘I’ is ‘It is neither true nor false that’. I take it the motivation for Williamson’s view is to offer a strategy for saving classical logic in the face of vagueness.
    But, suppose there are some propositions that are neither true nor false like “This book is blue.” Such propositions would be unknowable. Would the existence of such propositions present any problem for any plausible view of omniscience? I don’t see how.

    September 22, 2006 — 17:05
  • But, suppose there are some propositions that are neither true nor false like “This book is blue.”

    That view would not be epistemicism, though.
    The origal problem Mike presented isn’t about omniscience. It’s about whether God could create an object that is neither blue nor not blue. Such an object would entail a contradiction, on classical logic. One way to get around this is to be an epistemicist, and that then raises the omniscience worry.
    Other views of vagueness have to deny classical logic, and that’s a way out of the problem, but it leaves it less clear what it means to say that God can’t perform contradictions, because it’s less clear what counts as a contradiction. That was the context of this mini-debate (if we can even call it that).

    September 22, 2006 — 17:27
  • Christian

    I see. God cannot create an object that is neither blue nor not-blue.
    “Such an object would entail a contradiction, on classical logic.”
    And you think epistemicism entails that every object is blue or not-blue.
    So, if epistemicism is true, then God cannot create an object that is neither blue nor not-blue.
    But, if epistemicism is true, then there is some kind of omniscience worry. I don’t understand what this worry is I guess.

    September 25, 2006 — 12:30
  • Christian,
    The worry would be that, for some blue objects b, God might not know that b is blue. But this depends on a strong version of epistemicism. I noted above that Sorensen seems to endorse the position that some propositions are necessarily unknown (and so unknown by God).

    September 25, 2006 — 13:39
  • Christian

    Hi Mike,
    Great reference on the Sorensen paper! I read it, but I didn’t see anything in it that would imply there are obligations unknowable to God. For example, he gives a case where someone promises another to pay them a fair wage, but “fair wage” is vague, so from epistemicism it follows that one has an unknowable obligation. But, God would know what a fiar-wage is.
    He tries to cast doubt on an access principle that states that if we are obligated to do x, then we can know we are obligated to do x. All obligations are knowable. None of his counterexamples would imply that God has unknowable obligations though.
    So, I guess I still don’t see any problem for omniscience, not yet.
    “For some blue objects b, God might not know that b is blue.”
    If God is omniscient, then he knows for of all blue objects that they are blue and of all non-blue objects that they are not. If epistemicism is true, then (according to Jeremy), there are no objects that are neither blue nor not-blue. So, epistemicism and omniscience seem to me a happy pair.

    September 25, 2006 — 18:19
  • Chris,
    Sorensen argues, and he explicitly says–right off in his introduction–that he is going to argue, that some obligations are logically impossible to know. Didn’t you see that?? Now you might not be convinced that he has shown this, but that is an entirely different matter. My claim is that this is what he believes. Arguments aside, what I attributed to Sorensen is the position that some obligations are logically impossible to know (that is his language, not mine). You might want to revisit the article and confirm that this is precisely what he claims.

    September 25, 2006 — 20:23
  • Christian

    Hey Mike,
    Here’s the quote I think you’re pointing out:
    “(Access) If one is obliged to do x, then one can know one is obliged to do x.
    My negative goal is to refute access. Since `can know’ can be relativized to different background constraints, I advance counterexamples that range from practical impossibilities to logical impossibilities.”
    And I take it that since he says this you think that he is going to give an example of a proposition that is true, but is not logically possible to know.
    I can see why you think this, it is suggested right there above. It is not obviously entailed by what is said though. However, I don’t recall one example of a true proposition that is not logically possible to know. I’ll read again, sorry.

    September 25, 2006 — 21:19
  • For example, he gives a case where someone promises another to pay them a fair wage, but “fair wage” is vague, so from epistemicism it follows that one has an unknowable obligation. But, God would know what a fiar-wage is.

    God would not know of some cases whether they are fair wages.

    If God is omniscient, then he knows for of all blue objects that they are blue and of all non-blue objects that they are not. If epistemicism is true, then (according to Jeremy), there are no objects that are neither blue nor not-blue. So, epistemicism and omniscience seem to me a happy pair.

    In principle, yes. But proponents of epistemicism might tack on this extra bit that Sorenson explicitly adds. Williamson does not insist on it. That’s that vagueness is epistemic, meaning that the phenomenon of vagueness is not in what is true but in what can be known. The reason we can’t tell which borderline cases are blue and which are not blue is because it’s impossible to know, according to Sorenson. Williamson seems to me to be open to the view that it’s just impossible for us to know (but it’s impossible in principle for beings like us to know), whereas Sorenson seems to think it’s impossible in principle for any being to know. If he is right, then omniscience is not compossible with any vagueness.

    September 27, 2006 — 9:55
  • Christian

    “Sorenson seems to think it’s impossible in principle for any being to know. If he is right, then omniscience is not compossible with any vagueness.”
    This is interesting. I haven’t come across an argument for it yet, but I’m going to look up Sorensen’s stuff. He’s smart.

    September 27, 2006 — 10:55