Will anyone suffer forever?
March 15, 2011 — 10:51

Author: Josh Rausmussen  Category: Uncategorized  Tags: , ,   Comments: 12

Here’s a reason to think not.
Premise 1: A perfectly good God wouldn’t create someone who has a 99.999999999999999% chance of suffering forever. (I’m assuming here that the “suffering” is to an extent that makes one’s life not worth living.)
Suppose you have a dream in which an “angel” tells you that 99.999999999999999% of the people in Gabon, Africa will end up suffering in hell forever given their background culture and innate personalities. Do you believe it? Probably not, and not merely because you don’t believe God exists. You’d probably think this: “A good God wouldn’t permit there to be a person whose chance of escaping infinite suffering is so terribly slim.” That’s the intuition behind Premise 1.
Premise 2: If a perfectly good God wouldn’t create someone who has a 99.999999999999999% chance of suffering forever, then a perfectly good God wouldn’t create someone who has ANY chance of suffering forever.
Think about it this way. The difference between 99.999999999999999% and any other percent is FINITE, whereas the consequence is always INFINITE. How could there be a percentage that permits risking infinite suffering but a finitely different percentage that doesn’t? Or think about it this way. For every percentage p, either p is worth the risk of infinite suffering, or it is not the case that p is worth the risk. None of these terms are vague, so if Premise 2 is false, then there’s a p, such that a perfectly good God wouldn’t create someone who has p chance of suffering forever but would create someone who has a slightly smaller chance of suffering forever. Why should a slight difference in chance warrant an INFINITE difference in the conseqence that may be risked? It seems it shouldn’t.
Therefore: a perfectly good God wouldn’t create someone who has ANY chance of suffering forever.
Objection: what about the value of free will?

Of course, someone can object that certain important elements of free will would be taken away if no one had a chance of suffering forever–which is the natural and required consequence of ultimately rejecting the Good.
Notice that this objection doesn’t actually diagnose a problem with the premises of my argument. It seems to simply be an argument against the conclusion of the argument. Or, if this objection targets Premise 2, it doesn’t diagnose what’s wrong with the reasoning I gave in supoprt of Premise 2.
I reply with a series of questions:
1. Why couldn’t the required consequence of rejecting the Good be a finite stage of suffering followed by Annhilation?
2. Why couldn’t the required consequence of rejecting the Good be a finite stage of suffering followed by another chance?
3. Why couldn’t God consult counterfactual knowledge to create only those who would freely refrain from rejecting the Good?
5. What makes the freedom to be forever separate from God so valuable that it’s worth the risk of suffering forever?
These are just questions. Really, what I’d like to know is what’s wrong with my argument?

  • I can’t sign in for some reason. Really annoying.
    Anyway, really interesting post Josh! I’ll express two worries. First, chance is time-dependent. So, your chances of suffering eternally when born might be negligible, but very high when near death. God might know this and create you anyway, if (i) he can create beings that have an initially great chance of not suffering eternally and (ii) the reasons chances for suffering increase have entirely to do with your own free choices.
    Your premise (2) entails (I htink) that God could not (in premise (1)) create anyone with any positive chance of suffering eternally. Partly the reasoning has to do with Cantorian measures of infinity. We could conclude that it’s best not to use these measures, but some non-standard measures of infinities (hardly non-standard anymore, really).
    Like the post, in any case.

    March 15, 2011 — 19:34
  • Keith DeRose

    (I also couldn’t sign in for a while — and then, all of a sudden, I could.)
    I guess the best test scenario for your 2nd premise, Josh, would be one in which the chance of forever-suffering (in 2’s consequent) is extremely small, and there is a very high probability of forever-bliss. Though *I* get extremely risk -averse when forever-suffering is on the table, I imagine others might wonder why we should accept (2) if its consequent might concern such a scenario. One way they might press their worry is this: Compare the situation where you have the extremely good chance of suffering forever with the situation where you have only a very small chance of suffering forever and a very high chance of enjoying forever-bliss. Isn’t the latter a much better situation to be in? Wouldn’t you choose to be in the latter situation if you were given a choice between them? (“You can choose either having a very high chance of suffering forever, or a very small chance of suffering forever with a very high chance of enjoying bliss forever. Which do you choose?”) The latter seems a much better deal. (And I take the claim that one should choose the latter to be pretty solid, and something a theory of rational choice should accommodate — though of course tricky issues come up when infinities get involved.) But if it’s so much better for someone to be in the latter situation, what could be behind the conditional claim that if God couldn’t put someone in the first [worse] situation, He couldn’t put them in the second [better] situation?

    March 15, 2011 — 20:16
  • Josh:
    A. For a fuller argument, you should consider the possibility that the total amount of positive suffering in hell is finite rather than infinite. I don’t know how orthodox is the idea that it is in fact so, but at least the possibility seems logically possible. (Two scenarios: (1) the suffering decreases asymptotically in such a way that the sum/integral of total suffering is finite; (2) subjective time slows down so that in the nth objective year in hell, the sinner suffers 2^-n subjective years.)
    B. More seriously, it could be that the lower the probability of eventual damnation, the less free the agent is. Moreover, the value of, say, a year in heaven may be dependent on one’s causal history, in such a way that an agent who got to that state somewhat less freely is somewhat less well off during that year. And the “somewhat less well off” may add up to an infinitely lower value once one sums/integrates over the whole future. If so, then infinities are being played off against infinities, and it could well be, say, that a person with 0.5 probability of damnation could be worth creating since the 0.5 probability salvation that the person would get could have infinitely greater expected value than the expected value of the 0.9 probability salvation that a 0.9 probability of salvation person would get. (One can model this with non-standard analysis.)
    C. The following point should be a part of any good theology of hell:
    (*) Nobody suffers in hell so much that they would be better off being annihilated.
    There are several ways that (*) could be true and yet the suffering in hell be everlasting. Augustine’s metaphysics of evil as a privation could be true. Aquinas’ metaphysics of all existence as a participation in God could be true. It could be that the total suffering in hell is only finite and does not outweigh the good of human existence. It could be that the total suffering in hell is infinite but nonetheless there is enough of value in the life of the person in hell to outweigh it (either because of general facts about the value of human life or because of specific facts about values realized in that particular life). It could be that to receive just punishment is intrinsically good for the person suffering, in a way that overcomes the intrinsic disvalue of the harsh treatment that (at least partly) constitutes the punishment.
    Claim (*) is logically possible, if only because the finite amount of suffering scenarios are possible. Moreover, claim (*) is at least prima facie compatible with Scripture and Tradition. (The only Scriptural evidence I know against it is that it would have been better for Judas not to have been born. But to not have been born is not the same as to not have existence, since Judas’ existence preceded his birth–by about nine months I think, but almost everybody will agree that it preceded it by at least a few days. And significant portions of the Christian tradition–as exhibited by Augustine and Aquinas–are pretty much committed to (*).)
    I think that given (*), the conjunction of your premises 1 and 2 is significantly less plausible.
    D. How plausible premise 2 is may depend in part on what one’s reasons for accepting premise 1 are. Here is one kind of reason to reject premise 1. The only scenario worth taking seriously on which God creates a person who suffers everlastingly is where the person deserves the suffering for her freely and definitely chosen serious sin. But bracketing all question of eternal outcomes, it just seems very implausible that God would create a person who is 99.999999999999999% likely to freely and definitely come to be seriously culpable for serious sin. First of all, it’s not clear that it’s logically possible to have such a person, even bracketing the existence of God. It could be that as a necessary matter of fact, anybody who comes into existence with that high a probability of such sin lacks the freedom to be seriously culpable for that sin. Second, even if it were logically possible to have such a person bracketing the existence of God, any such person would either have to be in incredibly perverse circumstances or to have an incredibly perverse character, and it does not seem that God could put a person in such circumstances or give her such a character. But this line of thought does not license 2. For as soon as the probability gets around 0.5, then the worries that such a person couldn’t be sufficiently free, or that her character or circumstances would have to be very perverse, disappear.

    March 15, 2011 — 22:42
  • Joshua Rasmussen

    Very nice comments everyone.
    Mike: (i) let the chance be your initial chance (or else the result of multiplying all your chances across your life); (ii) Maybe the reasoning behind premise 2 can go through on non-Cantorian measures…
    Keith: good observations, but what about this?: if premise 2 is false, then there’s a cutoff value p, such that God wouldn’t risk the infinite consequence at probability p but would at probability p – .0000000000000001. Right? But, then, doesn’t it seem implausible that such an extraordinarily slim change in probability could make a difference with respect to risking an infinite consequence?
    (BTW, it was your paper “Universalism and the Bible” that first got me thinking on this track years ago.)
    Alex: Bravo, you found a way to completely undercut premise 2. But a few comments:
    1. I stipulated that the “suffering” in question was such that it makes one’s life not worth living. I had you in mind when I stipulated that! Given that stipulation, (*) is compatible my premises and so doesn’t call them into question. You should agree with the conclusion of my argument precisely because you accept (*), given what I mean by “suffering”. (Maybe there couldn’t be suffering like that, but that’s a different issue.)
    2. re B and D. These are clever and each undercuts my premise 2. However, I’m inclined to think that God could consult counterfactual knowledge to ensure that only those who would eventually freely repent are created. Given this background belief, the benefits you suggest could be achieved without risking there being people who experience endless suffering (to the not-worth-living extent). I realize that this move takes us to a different argument…

    March 16, 2011 — 11:49
  • Josh:
    Yes, we differ with regard to Molinism here. All the things I said about probabilities were predicated on the claim that God cannot make use of middle knowledge when deciding what to create (or maybe more weakly: cannot make use of middle knowledge of p→q when deciding whether to do something that brings it about that p). This claim is true if there is no such thing as middle knowledge or if there is but it must, to avoid circularities in the order of explanation, be bracketed in divine deliberation.
    I did miss the significance of the stipulation (which does make the title of your post slightly misleading). So, yes, I agree with you on your official conclusion, and so do Augustine and Aquinas and many other traditional anti-universalist theologians. Likewise, I accept Premises 1 and 2, because I think it is logically impossible for there to be anyone who would be better off not existing (whether because of the deep metaphysics of Augustine and Aquinas, or simply because a perfectly good being couldn’t allow such a state of affairs to take place).

    March 16, 2011 — 12:01
  • Joshua Rasmussen

    The title is perhaps misleading in another way, too, because the suffering I have in mind would add up to infinite suffering over infinite time, but of course, it’s possible for someone to suffer forever but only experience finite suffering over infinite time (e.g., if it were cut in half every thousand years). I’d clarify the title now, but it would throw off the comments…
    My impression is that most Christians who believe in endless hell do not think the lives there are worth living, but they should read your recent post. 🙂

    March 16, 2011 — 12:14
  • Josh:
    “My impression is that most Christians who believe in endless hell do not think the lives there are worth living”
    Anecdotally, you’re right. So you have a target here. Maybe not the great theologians of the Christian tradition, but many ordinary believers do. My hypothesis is that people take common descriptions of life in hell (fire and brimstone, etc.) and say that those descriptions are not descriptions of something worth living through. And then I say: either people are wrong about these things not being worth living through or the descriptions should be taken less literally (or both).

    March 16, 2011 — 13:43
  • Mike Almeida

    Hi Josh,
    I haven’t read all of the comments, but I did want to mention another worry about (2).

    If a perfectly good God wouldn’t create someone who has a 99.999999999999999% chance of suffering forever, then a perfectly good God wouldn’t create someone who has ANY chance of suffering forever.

    The argument for (2) seems to appeal to the expected disutility of hell, for any positive finite probability p of winding up there. So, the argument seems to go that for any assignments of finite postive probability p and q, we know that p(-oo) = q(-oo). So, God has as good a reason not to create anyone with p chances of going to hell as he has not to create anyone with q chances.
    But this is strange for a few reasons. Note that it is also true that, for any finite positive probability p and q of enjoying infinite happiness, p(+oo) = q(+oo). So, we have this response. For a p that approximates the infinitessimal, the expected value (for S) of creating S, Cs, is in (1).
    1. E(Cs) = p(-00) + (1-p)(+00) = undefined
    Whether we come out with a result that is infinitely positive or infinitely negative depends irrelevant features of the sum. If we are faced with a problem of this sort, we ought to rely on the probability distribution alone to decide what to do. God certainly would not be doing anything morally untoward in creating S in such a situation since S has a nearly certain chance of infinite happiness. God would be doing something wrong (maybe) if the probabilities were switched.
    2. E(Cs)= p(+00) + (1-p)(-00)
    Here he gives S a nearly perfect chance of eternal damnation.
    Last, I think it is misleading to put the argument in terms of chances. You make it sound like God is spinning a roulette wheel with your future. In fact, it is nothing like that. No matter what the ‘chances’ are of you going to hell, there is no one created by God who is such that it is not fully and completely in his control whether he goes to hell. It is the sense that such beings are not in control, but left ot chance, that (to me in any case) gives the argument force.

    March 16, 2011 — 16:04
  • Joshua Rasmussen

    The argument for (2) seems to appeal to the expected disutility of hell, for any positive finite probability p of winding up there.
    That was one thought. But another thought is that it seems strange that an extraordinarily tiny change in probability would make the difference with respect to God’s willingness to risk an infinite consequence. See my question to Keith. You might have a good answer.

    March 16, 2011 — 17:45
  • Joshua Rasmussen

    I’ve been thinking more about your points B and D. Regarding B, consider this question: would my son Micah’s relationship with us be enhanced if he had had a choice(s) between being ultimately separate from us (and then to suffer forever) or to be with us? It seems not. Indeed, to me it seems CLEARLY not. (And that seems true whether or not the suffering renders his live not worth living.) But then why would things be different in the divine case? I grant there’s a conceptual possibility that life in heaven is enhanced by our having made a significantly free choice not to be ultimately separated from God, but when I think about, this conceptual possibility seems really implausible. It’s one thing to be able to choose to reject a relationship; it’s a whole other matter to be able to choose to reject it forever. I might grant Micah the first capacity, but I certainly wouldn’t grant him the second. I perceive no value in that. Does God?
    Regarding D, it strikes me that perhaps the lower the probability the less responsible one is, but that no probability except zero would remove one’s responsibility altogether. If the cut-off probability between being responsible and not being responsible were some value v other than zero, then presumably it would be necessarily v. But it seems really implausible to me that it would be a necessary truth that v has a decimal value between 0 and 1: why that value rather than slightly higher or lower?
    So, while I think you’ve undercut my premise 2 at one level in the dialectic, the premise still seems plausible to me at the next level.

    March 17, 2011 — 8:25
  • I agree that any non-zero probability will include responsibility. But responsibility comes in degrees, surely, and lower degrees of responsibility are less valuable than higher ones, at least when it is responsibility for a good action.
    One might imagine the following variant on the parable of the prodigal son. The son who stayed behind could not conceive of leaving. He never chose to stay. The other son left and then chose to come back and stay. There is something more valuable about the state of the prodigal son at this point.

    March 17, 2011 — 15:21
  • jac

    Just passing by. On question 1,suffering followed by annihilation would be no suffering at all because of course, there could be no recollection of that suffering after the point of annihilation. Bad people of extreme fortitude-and perhaps a philosophical turn, might reason to that even if their actions were to be judged against them their suffering would be forgottenly transient. Also, they might unconsciousably reason that their victims will not suffer as any recollection of their suffering would need to be annihilated as a necessary condition of any elevated state. Or if not, then perhaps diminished by the value of the retrospective temporal finiteness to the infinite in contemplation.
    Annihilation would not only abolish meaningful suffering as the consequences of freely willed badness but also the loss of the prospect of reward for the Good.
    Perpetual hell may need to be a certainty although in the event it might not be.

    March 28, 2011 — 0:43