Open Theism and the Problem of Evil
December 10, 2008 — 8:35

Author: Alexander Pruss  Category: Molinism Open Theism  Comments: 20

It seems to me that some folks–perhaps not philosophers–think that Open Theism (OT) somehow significantly helps with the Problem of Evil. But I do not think it does. The natural reason to think OT helps is to say that if an omnipotent God foreknows that George will freely do some evil E, then God can prevent George from doing E, and OT means that God can’t foreknow it, so we can’t blame God for failing to prevent E. But this is confused. For it would be impossible for God to both foreknow–or even forebelieve–E and prevent E. Foreknowledge does let God put plans for an event into effect before the event happens, but for actual prevention of foreknown evils, what would be needed is Middle Knowledge, not foreknowledge.
I am curious if any philosophers have committed the error I criticize here.

Comments:
  • Mike Almeida

    The natural reason to think OT helps is to say that if an omnipotent God foreknows that George will freely do some evil E, then God can prevent George from doing E . . . But this is confused. For it would be impossible for God to both foreknow–or even forebelieve–E and prevent E.
    I’m not sure I see the confusion involved. Of course, there is no world in which God knows that S does A and also S does not do A. But that presents no problem. Suppose God knows that S will do A in w. God nonetheless can prevent S from doing A. Only were God to prevent S from doing A, it would be true that he never knew that S would do A. That is, were God to prevent S from doing A he would actualize a world w’ in which S does not do A and God knows that he does not do A. Here’s a possibly interesting consequence. There is no foreknowledge that does not reduce to middle knowledge. All that will happen is just all that would happen were God not to prevent it.

    December 10, 2008 — 11:33
  • Assume that classical theism is true, and thus some “closed” view of the future is true—either theological determinism, or Molinism, etc. Further assume that there is some world, w, such that God is considering creating w. If the classical definition of omniscience is correct, then it seems like God must, in this case, know all the future contingent truths about w. Surely, in this case, God could still prevent any evil event, E, in w from occurring by simply choosing not to create w. Instead God could create w*, where E never occurs. All God would need for this is Middle Knowledge. The OTist however denies that God can know, prior to creating w, what choices the creatures in w will freely make (i.e. OT denies Middle Knowledge). Thus, God could choose to create a world, w, where E might occur, without being to blame for the occurrence of E in w. The Molinist on the other hand supposes that God has Middle Knowledge, and knows that E will occur but does nothing to prevent E, while the theological determinist supposes that God truly foreknows all events, and allows E to occur. How doesthe OTist thinks that OT significantly helps with the problem of evil? On OT not only is it the case that God cannot foreknow E (on the hypothesis that E involves the choices of free creatures), it is also the case that God cannot have Middle Knowledge of E. Thus, the God is not, strictly speaking, responsible for E, just as I cannot be responsible for Alex’s freely choosing to eat dinner.
    I suppose someone might object that on the OT view God is, nevertheless, responsible for taking a huge risk. After all God creates a world not knowing what specific events will occur in it. But if we assume that God is essentially benevolent, and that God has infinite resources for adequately compensating those who suffer, this does not seem so bad.

    December 10, 2008 — 12:04
  • Mike Almeida

    Thus, the God is not, strictly speaking, responsible for E, just as I cannot be responsible for Alex’s freely choosing to eat dinner.
    But you can be responsible. I might be responsible for Smith freely doing X, even if I did not know (and could not have known) that Smith would do X. I might know that Smith has a gun and is angry and has shot people in circumstances like C before when angry, and I might put Smith in C. If God places me in circumstances in which I have harmed people before, those I harm might have a legitimate grievance against him. In a different context, I recall Clayton Littlejohn and Luke Gelinas advancing this line pretty persuasively on PB.

    December 10, 2008 — 14:51
  • Luke Gelinas

    I don’t remember advancing anything persuasively, but I do think that the open God will by any reckoning be responsible for permitting loads of actual evil.
    A.P., is the claim that the Molinist God would be responsible or culpable for moral evil? Even if the agents libertarianly freely commit the evil?
    Suppose OT is true, and that a friend is beaten and robbed. The open God allowed it. Why? Presumably it has something to do with some good, general or particular, God wants to realize. Suppose Molinism is true, and a friend is beaten and robbed. The Molinist God actualized a world where he knew this would happen. Why? The same general explanation: There was some good, general or particular, at stake, and God’s ways of realizing the goods, his creative options, are limited by the CCFs.
    I actually think that OTists have the tougher time here, since it seems more difficult for them to tell a plausible story in at least some cases about what justifies God’s non-intervention. They can’t, for instance, very well appeal to particular future moral goods–soul-making goods, say–that certain evils make possible, since it seems many times the open God won’t be justified to conclude much about which evils will produce what effects (the more so the more removed in time). Some OTists, I think Hasker in particular (though he’s not entirely clear on this), seem to think that the mere *possibility* that some evil will produce soul-making goods is enough to justify the open God. But this seems wrong in very many cases. Surely God wouldn’t be justified to permit leukemia to ravage a young child on the grounds that it *might* make possible future soul-making goods.

    December 10, 2008 — 19:47
  • Mike:
    It seems to me to be a conceptual truth that if one could do-X-on-account-of-a-belief-B, then it follows that it is possible for one to both have belief B and do X. Thus, if God could prevent S from doing A on account of God’s belief that S will do A, then God could both have the belief that S will do A and prevent S from doing A. But it is impossible that God could both have the belief that p and prevent p from occurring, since God’s beliefs are, necessarily, true.
    I do not think foreknowledge requires middle knowledge. But there seems to be less that can be done with the foreknowledge if there isn’t middle knowledge as well. (I have a difficult paper in a recent issue of F&P which explores this in respect of prophecy.)
    APT:
    A classical theism seems to require neither theological determinism nor Molinism. Classical theism seems compatible with simple foreknowledge, the view that, necessarily, God knows everything that will happen.
    Yes, for all w, God must know all the contingent truths about w. But one cannot then say that if God doesn’t like what is happening in w, he should just actualize w* instead of w. For that would require middle knowledge–God would have to know which world would result were he to strongly actualize such-and-such (if God creates a free creature, he may strongly actualize the creature, the creature’s character, the fact that the creature makes a free choice, but he does not strongly actualize the particular choice the creature freely makes). A non-Molinist, non-theologically-determinist God should not be thought of as choosing what world to actualize. Rather, he chooses what subset (this isn’t the best word–something like “skeleton” would be better) of a world to strongly actualize. What world would result from any given strong actualization is something he will not know, except perhaps in the particular case of the strong actualization that he in fact chooses.
    Your arguments show, at most, that the OT does better in respect to the problem of evil than the Molinist. But OT is not the denial of Molinism. OT is the denial of foreknowledge of future contingents. Moreover, what is most controversial about OT is not the denial of Molinism, but the denial of foreknowledge of future contingents. It is in the denial of foreknowledge, not in the denial of Molinism, that OT departs from the morally unanimous tradition of Western monotheism. Thus it is the denial of foreknowledge that is particularly in need of argumentative support. And the problem of evil does very little to suppose the denial of foreknowledge–it at most supports the much less controversial denial of Molinism.
    It is overkill to go for OT just to deny Molinism and theological determinism.
    That said, it is an interesting feature of the present philosophical landscape that folks like me, who (at least tentatively–I will, of course, abandon my view if it is shown to be inconsistent with the Church’s teaching) are neither Molinists nor theological determinists but hold on to simple foreknowledge, seem to be rare.
    LG:
    I think what one says to your view depends on whether one is more concerned about the evil deed as a moral evil in the evildoer or about the effect of the evil deed on the victim. If one is more concerned about the evil deed as a moral evil in the evildoer, following the Socratic and Christian tradition that it is much worse to do evil than to have evil done to one, then the OT or the defender of simple foreknowledge (SFK: foreknowledge without Molinism or theological determinism–I just made up the term) is in somewhat better shape than the Molinist. For given OT and SFK, it is certain that God cannot guarantee that significantly free choices are made and that they are all made correctly. If Molinism is true, it is possible (by Plantinga’s FWD) that God cannot guarantee that significantly free choices are made and that they are all made correctly, but it is very unlikely (the counterfactuals of free will have to all conspire together to ensure this, and that’s very unlikely). So with respect to the moral evil in the evildoer, OT and SFK are in better shape.
    Of course, the Molinist can bring in greater goods and all that, and I do no think there is much difficulty there. But it does take a bit harder work. OT and SFK, however, both agree that God cannot create significantly free creatures and guarantee they will act righteously.
    With regard to suffering evils done to one, or natural evils, the point you make about soul-building is a very nice one. You may be right that the Molinist is in a better position there than OT or SFK. On behalf of OT and SFK, however, I can say that the opportunity to have one’s soul built up is itself a great good, even if it is an opportunity one isn’t going to avail oneself of. (Interesting relevant side question: Is it good to to offer someone a good opportunity when one knows, as the Molinist God sometimes may know, that the person won’t accept it?)
    Mike (re. your second comment):
    I think we need to distinguish types of responsibility. In each of the following cases I have a responsibility for your being gored to death, where I knew you were an innocent:
    1. My ox gores you, but it has never gored anybody before, and I reasonably had no idea that it might gore anybody.
    2. I knew my ox gored people, but I let it go on the street anyway, and it gored you.
    3. I knew my ox gored people, and I knew that it would gore you if I let it go on the street, and I let it go on the street, and of course it gored you.
    4. I knew my ox gored people, and I let it go on the street in order that it gore you, and it did.
    I think there are at least three senses of responsibility corresponding to cases 1, 2-and-3 and 4 (maybe we want to break up 2 and 3, or maybe there is a continuum). It’s not a question of degrees–it’s a qualitative difference. In case 1, there is no wrongdoing along with the responsibility. But I still have a form of responsibility, as is clear from the fact that I had better try to make it up to your survivors, or at least to apologize (if I can do so in a manner that does not trivialize the situation). In case 4, I have certainly done wrong–I am a murderer. In cases 2 and 3, whether I have done wrong depends on further questions. I might have a defense based on the principle of double effect (PDE)–thus, I might have sent the ox out of my house to chase it away from my 12 children, whom it was trying to kill (and I didn’t have anything to kill it with). At the worst, in cases 2 and 3, I have not committed murder (except maybe by some laws), but something more like negligence or manslaughter.
    In respect of the problem of evil, the danger to the Molinist is of God’s potentially having the kind of responsibility in case 3. The danger to the OT and SFK defender is of God’s potentially having the kind of responsibility in case 2. The kind of responsibility in case 1 is not very problematic. So one strategy for the OT and SFK theodicist could be to argue down the responsibility from the sort in case 2 to the sort in case 1. I think Mike’s remarks suggest that this strategy is not going to be a success–inductive data shows God, on both OT and SFK, that there is serious danger of future evils. On the other hand, the Molinist and theological determinist does not even have the option to try to move down the responsibility type–he is stuck with responsibility of the sort in case 3.
    By the doctrine of double effect, only the kind of responsibility in case 4 is going to be absolutely incompatible with perfect moral goodness. The other kinds of responsibility can permissibly be undertaken with good enough reason. So the question for the theodicist is going to be to argue that there is good enough reason. This will be somewhat harder in the case of Molinism and theological determinism than OT and SFK. The theological determinist has it hardest here. But none of this is a knock-down argument against Molinism or theological determinism.

    December 10, 2008 — 22:44
  • Luke Gelinas

    I’m not sure I follow completely. I guess I’d need to hear more about what reason there is to think it false that the CCFs are such that every world as good as ours contains at least as much moral evil as ours.
    Even if it’s certain that the OT/SF God cannot guarantee that creatures go right, the OT God (at least) actually does permit them to go wrong when he could prevent it (if OT is true). I was thinking that the open God’s permitting some moral evil E might in some cases be morally equivalent to the Molinist God’s weakly actualizing E. Is your claim that OT and SF have the advantage b/c it’s just implausible to think (given the amount of evil in the world) that the CCFs constrain God this badly?
    I’m not sure what I think about the mere opportunity for soul-making–it might be good, but I’m not sure it’s a *great* good. I’m not sure how much innocent suffering would be justified by it. A world in which I consistently have opportunities to respond virtuously to some fairly intense pain of yours (a bad toothache, say) but never do so doesn’t seem as good to me as a world with no bad toothache (all else equal).
    I like the hierarchy of responsibility. But suppose an OTist proposed a replacement for (2):
    (2*) I knew my ox was prone to be a surly fellow at times, but I didn’t know just how violent he could be; I’m surprised that my ox gored you so bad; I didn’t think he was capable of *this*.
    Does this change things at all? I think it might, though not strictly speaking with respect to God’s culpability.

    December 11, 2008 — 0:08
  • Mike Almeida

    It seems to me to be a conceptual truth that. . . if God could prevent S from doing A on account of God’s belief that S will do A, then God could both have the belief that S will do A and prevent S from doing A.
    Alex,
    This is closer to a conceptual truth: God could prevent S from doing A on account of God’s belief that otherwise S will do A, then God could both have the belief that S will otherwise do A and prevent S from doing A.
    And it is certainly possible for God to believe that S will do A unless God prevents it and also to prevent S from doing A. This is why I suggested that the foreknowledge that S will do A is in fact a kind of middle knowledge. It is the knowledge that were God not to prevent S from doing A, S would do it.

    December 11, 2008 — 9:04
  • Mike Almeida

    I don’t remember advancing anything persuasively, but I do think that the open God will by any reckoning be responsible for permitting loads of actual evil.
    I take back my observation that Luke made a persuasive argument at some point about something…. 🙂

    December 11, 2008 — 9:13
  • Luke Gelinas

    I knew you must be confusing me with someone else.
    I’m not sure I understand the SF view. What exactly is it that God foreknows, and when? God foreknows the history of the world, right? But is it only after (temporally or logically) God creates the skeleton of the world that he foreknows its history? I take it from Alex’s comments that there’s not supposed to be multiple worlds whose history God foreknows before he creates; is that right? If there were, it seems there would at least be true CDFs, counterfatcuals of divine freedom, and that God would know them. God would know that were he to actualize skeleton world W, W’s actual history would follow.

    December 11, 2008 — 9:36
  • Mike:
    The “unless” had better be have a semantics that isn’t merely inherited from the material conditional for this to work. It had better be a genuine counterfactual, or something like that. And then we need Molinism.
    So our disagreement is not so much over whether Molinism is needed for such preventions, but over whether Molinism is needed for foreknowledge.

    December 11, 2008 — 11:58
  • Mike Almeida

    So our disagreement is not so much over whether Molinism is needed for such preventions, but over whether Molinism is needed for foreknowledge.
    I’m not sure I see why you’re worried. The ‘unless’ is embedded in God’s belief in (1).
    1. God believes that [S will do A or God prevents S from doing A].
    2. God prevents S from doing A.
    (1) and (2) are consistent. In the closest world in which God prevents S from doing A he believes that he prevented him from doing A. Now suppose that S will actually do A.
    3. S will do A.
    (1) is also consistent with (3). (1) & (3) hold in the actual world, and (1) & (2) hold in the closest world in which God prevents S from doing A.

    December 11, 2008 — 13:45
  • Mike Almeida

    Luke,
    All kidding aside, the argument here http://prosblogion.ektopos.com/archives/2007/01/more-thoughts-o.html was pretty strong, I think. But there was a lot of argument involved. Later on Clayton argued along similar lines here http://prosblogion.ektopos.com/archives/2007/11/moral-responsib.html. For what it’s worth, I mention those arguments (giving credit where its due) in an upcoming paper.

    December 11, 2008 — 13:57
  • Luke Gelinas

    Mike,
    Thanks. Though I think I wanted to frame the advantage for the open God in such a way that it didn’t imply that the Molinist God would be culpable for anything. I want to carve out a way of saying that, while the Molinist God does nothing wrong, he still can’t realize these other goods that the open God can. (I take it this is different from A.P.’s approach above.)
    I’m still working on the paper that descended out of this discussion (I give it at Baylor in February). Yes, I recall some heated and prolonged discussion in that thread!

    December 11, 2008 — 15:21
  • Hi Alex,
    I think some theologians may be confused in the way that you imagine, but as far as I can tell, not any philosophers. In fact, William Hasker has repeatedly made the point that simple foreknowledge wouldn’t enhance God’s providential control over what happens.
    I think the main reason OTers like me hold that OT is better on the problem of evil is that it gives a better response to Rowe-type evidential arguments. The most popular response, skeptical theism, says that even for the most *apparently* pointless (gratuitous) evil (you know, burn-injury Bambi) there must be a good that logically requires it, even though it is utterly inconceivable how there could be such. Many have argued that this is too skeptical, and really more skeptical than the theist is honestly willing to be. Would we have to be opened minded about, e.g. Hitler – after all, we can’t be certain that God knew, and even told Hitler, some good, inconceivable to the rest of us, that logically required the Holocaust. But that’s absurd – we know those actions were evil, and we ought not be skeptical.
    So OT allows the theist to say, as van Inwagen, Hasker, and others have argued, that theism is consistent with there being gratuitous evils. But then, Rowe has no argument anymore – now apparently gratuitous evils needn’t count as evidence against theism.
    Best,
    Dale

    December 12, 2008 — 7:41
  • Dale:
    I don’t see how it is specifically OT that lets you say these things about Rowe-style arguments. It seems that you get to say these things because of the denial of middle knowledge and theological determinism. While that you can say these things may count somewhat in favor of OT, it does not count in favor of the most controversial doctrine of OT, namely the denial of foreknowledge–at least if you agree with Hasker’s point.
    In any case, it seems to me that the OT in these kinds of cases is still denying that these are gratuitous evils, at least if one understands a gratuitous evil as one which a perfectly good, omniscient and omnipotent being would have no justification in allowing to happen. That one doesn’t know that E is going to happen is a pretty good justification for allowing E to happen!
    As for Hasker’s point, it’s close to the truth but not exactly right. For a God who has simple foreknowledge can make use of this foreknowledge to react to evil choices before the choices are made, for instance setting into motion a chain of natural events that counteracts the evil after it happens. (A deity without foreknowledge could only react as soon as the choice were made, and thus could not prepare the world for that evil.) Moreover, a God who has simple foreknowledge can issue prophecies guaranteed to be true (see my recent F&P paper on prophecy and foreknowledge).

    December 12, 2008 — 8:43
  • Luke Gelinas

    Hi Dale,
    I don’t remember there being any point in van Inwagen’s vagueness defense of pointless evil where he explicitly leans on OT premises. In his latest book, at least, I seem to recall him going out of his way not to do so. Van Inwagen might in fact affirm an open future, but is there anything in his defense of gratuitous evil that requires it?
    There is also the existential problem of evil to consider. OT might help here, in a way more traditional views can’t. I think both Boyd and Sanders say stuff like this in places.

    December 12, 2008 — 10:05
  • patrick todd

    Hi Alex,
    You say, “In any case, it seems to me that the OT in these kinds of cases is still denying that these are gratuitous evils, at least if one understands a gratuitous evil as one which a perfectly good, omniscient and omnipotent being would have no justification in allowing to happen.”
    But my impression was that gratuitous evils in the sense at stake were evils which are not necessary for some greater good or the prevention of some evil equally bad or worse. At any rate, that’s what the dispute is between Hasker/PvI, and Rowe.
    Has anyone read Hasker’s new book, *The Triumph of God over Evil*? Therein, and other places, if I recall, Hasker argues that we should reject Rowe’s thesis that there cannot be gratuitous evils (in the above sense). Instead, he argues that the very reason why we are supposed to fight against such evils is precisely *because* they are gratuitous — that, unless we prevented them, the world would be worse off. In other words, if God guaranteed that there were no gratuitous evils in Rowe’s sense, we would have no good reason to fight them. But this would undermine the soul-making goods made possible by our fighting evil. (At any rate, that’s how I recall it going — unfortunately, I can’t look at the book, since it and some of my other books got stolen over the summer!)
    Now, Hasker goes on to argue that, on Molinist suppositions, God has to arrange the world so that every evil he allows is not gratuitous. Same with theological determinism. The only model that allows him not to do this is open theism. One *might* suppose that simple foreknowledge would too, but he argues that, with respect to providential control, simple foreknowledge and open theism are identical. So it can’t help here.
    So, FWIW, here’s how Hasker thinks OT helps with the problem of evil. Gratuitous evils are necessary for the right sort of soul-making. Only OT allows for gratuitous evils. So only OT can get the right sort of soul-making. But that sort of soul-making is an important part of a response to the problem of evil.

    December 12, 2008 — 16:20
  • Patrick – thanks for the precis. I need to have a look at that book. Sounds like Hasker has developed his views beyond what I’ve read.

    December 13, 2008 — 6:48
  • Luke Gelinas

    I don’t understand why Hasker’s general argument requires OT. Why must the Molinist God arrange things so that there’s no pointless evil? On Hasker’s own premises, wouldn’t the Molinist God see that, were he to do so, important soul-making goods would be lost? And that therefore a better world would be one which included gratuitous evil?
    At any rate, Hasker’s argument rests on some tricky premises. All that God needs to ensure that humans stay properly motivated is that we *believe* that there are pointless evil. But God might be able to permissibly see to it that we do so, even if no evils are pointless. (In fact, God might have to do nothing, or very little, on this score. How many people do you know who think that each and every evil is nec. for the creation of some greater good, etc.? Other than analytic philosophers of religion, at any rate. And even someone who did hold this belief would need to connect the dots in the way Hasker does before it posed a threat to their motivation.)
    Moreover, I think Hasker’s argument might need something like a patient-centred requrement on the justification of suffering. Suppose some evil e results in 10 units of suffering for Joe, but is necessary for the creation of 20 units of good in the world, none of which redound to Joe. It seems I could still be motivated to prevent Joe’s suffering, even if I believed it wasn’t pointless. My grounds would be that even if the evil doesn’t make the world worse on the whole, it still makes *Joe* worse off. And that would be enough to motivate me to prevent it.

    December 13, 2008 — 8:57
  • Dan Speak

    Great discussion, folks. I just want to add that David Hunt has an excellent paper in IJPR in which he considers some of these issues. I think it is called “Evil and Theistic Minimimalism”. David has also argued in other contexts for the usefulness of simple foreknowledge for some aspects of divine providence (contra some of the gestures in the original post)
    For what it is worth, I side with those who think open theism is going to help the problem of evil.

    December 30, 2008 — 21:53