Virtual Colloquium: Timothy Perrine, “Skeptical Theism and Practical Reasoning”
March 24, 2017 — 6:00

Author: Kenny Pearce  Category: Existence of God Problem of Evil  Tags: , , , ,   Comments: 15

Today’s virtual colloquium paper is “Skeptical Theism and Practical Reasoning” by Timothy Perrine. Perrine is a graduate student at Indiana University, where he is finishing his dissertation “Epistemic Value and Accurate Representation.” He works primarily in epistemology and philosophy of religion while dabbing in other fields. Some of his work has appeared in journals like Synthese and Faith and Philosophy as well as several edited volumes.


Skeptical Theism and Practical Reasoning

Timothy Perrine

Skeptical theism is an important and popular response to arguments from evil. Skeptical theists urge a kind of skepticism about our ability to discern the possible reasons God might have for permitting the evils we observe. They then propose general epistemic principles concerning when an interference is reasonable or it is reasonable to believe something is evidence. By combining their skepticism with such epistemic principles, skeptical theists aim to undermine arguments from evil.

But skeptical theism is not without its critics. Many critics allege that its skepticism leads to other skepticisms that are problematic. It is useful to have a taxonomy of the alleged skepticisms. Some critics allege skeptical theism leads to non-moral skepticism and others moral skepticism. Among non-moral skepticisms, critics urge that skeptical theism leads to skepticism regarding the external world or God’s commands. Among moral skepticisms, critics urge that skeptical theism leads to skepticism regarding the rightness/wrongness of action, all-things-considered value, or practical reasoning.

In this paper, I will be focusing exclusively on the objection that skeptical theism leads to skepticism regarding practical reasoning. Put crudely, that objection alleges that skeptical theists cannot reasonably conclude that they should prevent evils. For skeptical theists claim that there is a good that justifies God’s permission of an evil. But they also claim that we shouldn’t expect to see what that good is. Thus, even if a skeptical theist could easily prevent an evil for which she cannot see any outweighing good, she should not reasonably conclude that she should prevent it. For, though she cannot tell how, she thinks it would be best all-things-considered to allow both the evil and its justifying good to exist. But—the thought goes—such reasoning is problematic and so, by extension, is skeptical theism.

The aim of this paper is to respond to this objection. In section I, I briefly review skeptical theism and articulate a distinction between two kinds of God-justifying goods. Some goods justify God’s permission of an evil because the existence of the evil is necessary for the existence of the good and the good outweighs the evil. Some goods justify God’s permission of an evil because God’s permission of the evil is necessary for the existence of the good but the existence of the evil is not necessary for the existence of that good.

In section II, I develop this objection from practical reasoning skepticism at greater length, paying particular attention to an influential defense of it due to Michael Almedia and Graham Oppy. In section III, I argue that Almedia and Oppy’s defense of this objection fails because it runs afoul of the distinction between two kinds of justifying goods.

In sections IV and V, I am proactive, sketching a way that a skeptical theist might think about her skepticism. I argue that when deliberating a skeptical theist might in effect reasonably ignore her skepticism regarding access to justifying goods. For her skeptical theism by itself rarely gives her any reason for thinking it would be better or worse to permit an evil. And if she can reasonably ignore her skeptical theism when deliberating about whether to prevent an evil, she can reasonably arrive at the same conclusion that non-skeptical theists do, namely, that she should prevent the evil. In this way, skeptical theism need not lead to practical reasoning skepticism.


The complete paper is available here. Comments welcome below!

Comments:
  • Hi Timothy,

    I’d like to make a brief point about the commitments of skeptical theism, at least as a reply to non-theistic arguments from evil. You say that

    A couple of comments about ST1-3. First, ST1-3 do not claim that we have a good reason for thinking that the goods we know of aren’t representative of the possible goods that there are. Rather, ST1-3 are more agnostic than that: we lack a good reason for thinking they are representative; it may be that they are, it may be that they are not.

    However, it seems to me that a theist who uses skeptical theism as a reply to nontheistic arguments from evil is committed to the claim that they aren’t representative (when it comes to the property of justifying God’s permission of evil.), or at least that they almost certainly aren’t representative. Otherwise, they would need a different reply to the non-theistic arguments from evil, rendering the skeptical theist reply insufficient to justify their theistic stance.

    A nontheist might argue as follows:

    P1. Either the goods we know of aren’t representative of the possible goods that there are with respect to the property of justifying God’s permission of evil, or they are.
    P2. If they are, then God does not exist (because of the argument from evil the nontheist proposes).
    C. Hence, if God exists, they are not.

    Given that the theist is committed to the conclusion that God exists, she seems committed to the conclusion that they are not representative, or else she’d have to reject P2, which would require that she addresses the argument from evil without resorting to skeptical theism.

    A probabilistic argument can be given as well, like:

    Q1. Either the goods we know of aren’t representative of the possible goods that there are with respect to the property of justifying God’s permission of evil, or they are.
    Q2. If they are representative, the probability that God exists is less than 0.1.
    Q3. P(God)=P(God|Representative)P(Representative)+P(God|¬Representative)P(¬Representative)=< 0.1P(Representative)+(1-P(Representative))=1-0.9P(Representative).
    C: P(Representative)0.9 to God’s existence, it follows that P(Representative)<(1-0.9)/0.9=1/9.

    It seems the theist ought to reject Q2 by arguments other than skeptical theism, or else assign probability less than 1/9 to the hypothesis that the goods are representative (She might reject making probabilistic assessments about God's existence, but I think that's a mistake – i.e., she is making implicit probabilistic assessments when she asserts that God exists, even if she rejects the explicit probabilistic claim).

    March 25, 2017 — 13:19
  • Tim Perrine

    Hi Angra,

    Thank you for the comment. I’m not sure I fully understand why you think that skeptical theists can’t be as agnostic as I claim they can in the paper. I was wondering if you’d be willing to say more. Here are two ideas as to how I’m thinking about things that might help identify a possible locus of disagreement.

    First, I think skeptical theists can allow for there to be cogent responses to argument from evil that don’t turn on their skeptical theism. To take a simple case, I personally think that there are at least three distinct problems with J. L. Mackie’s 1955 argument from evil, none of which turn on issues involving skeptical theism. Further, I don’t think skeptical theism’s “skeptical theses” like ST1-3 are necessarily sufficient to respond to arguments from evil. They usually must be combined with principles in epistemology concerning evidence, reasonable inference, or the like to get purchase against other arguments. Finally, I’m not sure ST1-3 are the right skeptical theses to respond to all arguments from evil. I think of skeptical theism as a family of views, which get particular content due to the particular arguments from evil they are criticizing. Thus, even if ST1-3 aren’t quite right for responding to a particular argument, there might be some other set of skeptical theses that are similar to them that might do part of the job.

    Second, consider an argument against theism that requires the truth of:

    (R) The goods we know of are representative of all possible goods when it comes to the property of serving as a God-justifying good.

    For such an argument to be good—e.g., to have true premise, utilize valid or inductively strong inferences and the like—(R) must be true. But for it to be an argument that we can reasonable accept it must also be that it is reasonable for us to accept (R). After all, if we are not in a position to reasonably accept (R), and some argument requires it, then we are not in a position to reasonably accept the argument as a good argument. Thus, for the skeptical theist to criticize such an argument it is sufficient for her to criticize the reasonability (R). Further, she can criticize the reasonability of (R) without having to say that (R) is false or that it is reasonable to believe that (R) is false. If the appropriate attitude towards (R) is just agnosticism, then that should be sufficient to undermine any argument that requires (R).

    For this reason I do not think a skeptical theist has to say that something like (R) is false in order to respond to some particular argument from evil. This is why I’m skeptical of your ‘P2’ and ‘Q2. But you don’t specify a particular argument from evil. Perhaps you have one in mind that requires the skeptical theist to hold that she must reasonably reject (R) as false instead of being merely agnostic towards it?

    Best,
    Tim

    March 25, 2017 — 17:04
    • Tim,

      I’m not sure I understand why you don’t find my previous argument persuasive. It doesn’t rely on (R).

      Perhaps the following variant will explain my objection more clearly:

      Let RAFE be any random argument from evil, usually made under the assumption (R) that you mention, and which (under (R)) reaches the conclusion that God does not exist, or at least that the probability of God’s existence is less than 0.05 (for example; most arguments from evil would implicitly or explicitly give a lower probability than that to God’s existence).

      We can modify RAFE as follows:

      P1: The probability of R is no less than 0.4.
      C1: Hence, the probability of ¬R is no more than 0.6.
      P2: Because of RAFE, if (R) obtains, then the probability that God exists is less than 0.05.
      C: P(God)=P(God|R)P(R)+P(God|¬R)P(¬R)<0.05P(R)+P(¬R)=0.05(1-P(¬R))+P(¬R)=0.05+0.95P(¬R)=<0.05+0.95*0.6=0.62

      So, from those premises, the conclusion is that probability of God is less than 0.62. But the skeptical theist claims that God exists, so it seems to me she implicitly is committed to rejecting that the probability is less than 0.62.

      So, she ought to reject at least one of the premises, P1 or P2.
      Now, if she rejects P2, then she may not use skeptical theism or the principles in question to justify that rejection, since whether those premises are true has no impact on whether assuming (R) obtains, RAFE shows that the probability that God exists is less than 0.05.
      So, if she rejects P2, she ought to give a reply to RAFE that is independent of skeptical theism.
      Moreover, given that RAFE was a random argument from evil with the only condition that it reaches the conclusion or implies that the probability of God’s existence is less than 0.05, then she ought to reply to all arguments from evil that meet that condition by means of a reply independent of skeptical theism.

      The alternative is for her to reject P1, but then she’s implicitly committed to assigning probability greater than 0.6 to ¬R.

      March 25, 2017 — 20:47
      • Tim Perrine

        Hi Angra,

        Thanks for the follow up! I think I have a better understand of your points. In response, I’ll say this.

        First, skeptical theists who are theist will want, as you point, the probability of theism to be close to 1. But they will want that only in the final analysis. It is perfectly fine and consistent for them to permit that, given some evidence, theism does not have that probability. So I think a theist who is a skeptical theism could, in effect, accept your “Modified RAFE” and its conclusion that the probability of God is .6. Further, she need not object to that argument. What she does need to do is hold that given other evidence or arguments—e.g. fine tuning argument, religious experiences, etc.—that the probability of God shoots back up to closer to 1. So she could be perfectly consistent and not reject any premises or inferences in the argument you give. (I can say more about this, and the role of skeptical theism, if you want.)

        Second, focusing on P2 of your Modified RAFE, I take it to be equivalent to (more or less): if R obtains, then given the premises of RAFE, the probability of theism is below .05. Your point which I don’t think I understood initially is that if a skeptical theist wants to reject that premise she must do so for reasons independent of skeptical theism. I think you’re probably right, given some caveats that seem unimportant to your argument. But four subpoints. (1) That does not strike me as problematic. As I already mentioned, there can be problems for arguments from evil that don’t turn on skeptical theism. (2) You haven’t actually formulated an argument for evil. So whether or not a skeptical theist should reject P2 is underdetermined. You’re right that IF she rejects P2 it should be for reasons other than skeptical theism; and IF she does not, it is presumably because it is fine assuming R. But none of that shows what she should do until we know what exactly the particular RAFE is under discussion. As a corollary (3), we cannot generalize that skeptical theists ought to respond to all arguments for evil by means independent of skeptical theism. If we don’t know that a skeptical theist should reject P2, then we can’t generalize off the assumption that she should. Finally, (4), a formal point. If we understanding P2 in the way I’ve specified, the conclusion of the Modified RAFE does not follow. What follows from P2 is:

        P(Theism/R obtains and the premises of RAFE) = .05

        But it does not necessarily follow that:

        P(Theism/R obtains) = .05

        As you seem to assume when giving assignments to your conclusion. Now this later thing might follow from the former if all of the premises of the RAFE had a probability of 1. But again we’d have to look at a particular argument to know whether that was plausible.

        Finally, I think that many skeptical theists would reject your P1. But it does not follow that they are required to assign a probability of .6 or greater to ~R as you say. For instance, they might reject P1 because they simply have no idea what probability to give R. An analogy, consider (L): there is sentient life in the universe that is non-human. Suppose you assign a .4 probability to L. I might reasonably reject that assign, not because I think its really .2 or .75 or whatever. Rather, I might just be in the dark about L and find any particular assignment misrepresents my state of ignorance.

        You might object that if skeptical theists claim to be in the dark about R then they cannot assign a high probability to theism. After all, P(T)=P(T/R)*(R) + P(T/~R)*(~R), by the total probability theorem. But I think this point is incorrect for points made in the first comment. Even if the theist is unsure about the probability of theism considered in isolation, she might be much more confident given other bits of information. (Again, consider the analogy. Let (S) be ‘I am sitting right now.’ Given the total probability theorem, we get P(S)=P(S/L)*P(L) + P(S/~L)*P(~L). But I don’t think it is particularly plausible that my ignorance of L means I can’t assign a high probability to S, given everything I know and am experiencing.)

        Best,
        Tim

        March 26, 2017 — 10:16
        • Tim,

          Thanks for the follow-up as well!

          In re: final analysis vs. intermediate stages, I agree, but my point was about what the skeptical theist was committed to, so if she’s committed after her the final analysis to giving very low probability to R, that works for me as well.

          On the issue that I haven’t formulated an argument from evil, the choice of a random argument from evil was deliberate, because the point is that the skeptical theist ought to reject every single one of them (i.e., among those who give in the end less than 0.05 to God’s existence, which is at least nearly all arguments from evil) on grounds other than skeptical theism, unless she rejects P1, or even Q1 below.

          So, if it turns out that in some cases, she does have replies that don’t rely on skeptical theism, she’s still committed to giving low and even very low probability to the hypothesis that the goods we know of are representative of the goods there are with respect to God’s reasons for allowing them, as long as there is at least one argument from evil that she only rejects on the basis of skeptical theism.
          On the other hand, if she rejects every single argument from evil on a basis different from skeptical theism, skeptical theism isn’t doing much (or any) work, it seems to me.

          I don’t understand the following: You say

          You’re right that IF she rejects P2 it should be for reasons other than skeptical theism; and IF she does not, it is presumably because it is fine assuming R.

          But if she does not reject P2, then she rejects P1 (and even Q1), so she’s not assuming R, but rather, committed to R’s having low probability, or even very low with the Q1 variant.

          Regarding your (3) point, I do not claim that she ought to reject P2. Rather, I claim that she ought to reject P2 unless she rejects P1 (and the same for Q2 and Q1). My goal was not to show that the skeptical theist should reject any of the arguments from evil on the basis of something other than skeptical theism, but that unless she does that for every single argument from evil (well, at least nearly every one), she is committed to assigning low – and even very low – probability to the hypothesis that the goods we know of are representative of the goods there are with respect to God’s reasons for allowing them.

          In re: your formal point, RAFE argues from accepted premises (i.e., some instances of moral evil, suffering, etc. that we know happen), and conclude that the probability of God is less than 0.05, so I included that info the background information to simplify notation (whenever one makes an assessment about the probability of God, that’s on the basis of some accepted background info, which is implicit and not added as P(God|accepted background info)).
          If you want to specify that it’s on the basis of some premises, of course that’s correct and gives more details. But as long as those premises are accepted by the skeptical theist (and they usually are; we’re talking about generally accepted instances of moral evil or suffering) that does not affect the central point I’m making.

          March 26, 2017 — 12:34
          • Tim Perrine

            Hi Angra,

            Thanks again for the follow up. I’m a little embarrassed because I feel like I’m missing something, but I want to see if I can articulate your line of thought. I take your overall argument to be something like this:

            (1) For all the arguments for evil, either (i) the skeptical theist rejects them all for reasons having nothing to do with skeptical theism, (ii) she rejects all of them for reasons having something to do with skeptical theism, or (iii) she rejects some for reasons having to do with skeptical theism, some for reasons not having to do with skeptical theism.
            (2) If (i), skeptical theism doesn’t seem to be doing any work.
            (3) If either (ii) or (iii), then she is committed to (R) having a very low probability.
            (4) Therefore, skeptical theism doesn’t seem to do any work or the skeptical theist is committed to (R) having a very low probability.
            (5) Therefore, if the skeptical theist is not committed to (R) having a very low probability, then her skeptical theism doesn’t seem to be doing any work.

            Now I want to hold a position like (iii) while also maintaining that the skeptical theist can be agnostic towards (R) or something like it. So I guess to be consistent I’d have to reject the third premise (3).

            The reason I sketched for thinking the third premise was false was this. For any argument that assumes (R) implicitly or explicitly, a sufficient response to rejecting that argument is to be agnostic towards (R). Since one can be agnostic towards (R) without assigning it a very low probability, the skeptical theist can endorse positions (ii) or (iii) without being committed to (R) having a very low probability. (Perhaps what I said at one point was confusing. I didn’t mean to suggest that a skeptical theist could be agnostic about (R) while also assuming it. I meant that if she does not reject P2 from above, this is because she’s willing to concede assuming (R)—i.e. the antecedent of P2—the rest follows. Of course, she need not actually assume (R) to agree to a conditional that has (R) as the antecedent.)

            As you’ve pointed out, we can always reformulate an argument from evil to make explicit assumptions like (R). (This is what you did in your first two posts.) But I think this is consistent with the point of the previous paragraph. To return to the analogy with non-human life in the universe, for any argument that assumes that there is not any non-human sentient life in the universe we can always create a different argument that makes such an assumption explicit. But if I skeptical with the first argument because of that assumption, then making it explicit shouldn’t fundamentally change my mind.

            Let me try one more thing. Anytime P entails Q but not conversely P(P) > P(Q). Suppose ~R entails theism but not conversely. Then P(~R) > P(T). Further, suppose one wants to hold that P(T) = .95. Then one will have to hold that P(~R) > .95 and by corollary P(R) < .05. This would be an argument that any theist—include skeptical theists who are theists—must hold that R has a low probability. But the problem is that I don’t think that ~R entails theism. I think one can be an atheist or agnositic or whathaveyou and accept ~R.

            Anyway, I’m sorry if I’m missing something obvious.

            Best,
            Tim

            March 27, 2017 — 10:16
  • Addition.
    Let’s say that the skeptical theist assigns probability no less than 0.9 to God’s existence. Then one might argue as follows:

    Q1: The probability of ¬R is no more than 0.89.
    Q2: Because of RAFE, if (R) obtains, then the probability that God exists no more 0.05.
    C’: P(God)=P(God|R)P(R)+P(God|¬R)P(¬R)=<0.05P(R)+P(¬R)=0.05(1-P(¬R))+P(¬R)=0.05+0.95P(¬R)=<0.05+0.89*0.95=0.8955<0.9

    Then, the skeptical theist seems committed to rejecting Q1 (and not just R), unless she has a reply to nearly all arguments from evil (i.e., all of the arguments from evil implicitly or explicitly committed to the claim that the probability of God's existence is less than 0.05) that is independent of skeptical theism + the principles of epistemology you mentioned.

    March 25, 2017 — 20:55
  • Hi Tim!

    Thanks for sharing your paper. I’ve actually been discussing, recently, the distinction between (a) an evil being justified since it is required for a good, and (b) the permission of an evil being justified since it is required for a good. I have moved from thinking there is a real difference behind it to being totally unsure. I worry that this distinction might be the product of mis-describing the relevant goods and evils in certain cases, and their relations. So perhaps you can help me out with this.

    Consider your own example of a child who is permitted a painful public embarrassment, due to her lack of practicing his clarinet, for the sake of a lesson in responsibility. You say:

    “The evil itself—the painful public performance—is not necessary for the good of developing responsibility. Rather, the parent’s permission of that evil is necessary for the good of developing responsibility; after all, if the parent had forced her child to practice the clarinet, the child would not have had the opportunity to learn responsibility for his actions.”

    In this example, it seems that the *evil* is the painful public performance and that the *good* is the lesson in responsibility. You say that here the evil is not necessary for the good, though the father’s permission of the evil is. Why is the permission necessary? Because “if the parent had forced her child to practice the clarinet, the child would not have had the opportunity to learn responsibility for his actions.”

    My worry is that a similar counterfactual is true of the relationship between the evil and the good as well. So if such counterfactual obtaining is your reason for claiming that the permission is necessary, it seems you should also claim that the evil is necessary. Consider: Had the child not experienced the painful public performance, the child would not have learned responsibility for his actions. If Johnny’s dad permits him to choose to play video games instead of practicing the Clarinet, hoping that this will provide “the opportunity to learn responsibility for his actions”, but Johnny’s failure at the recital is met with utter indifference by Johnny (perhaps even some amusement, since he truly does not care), then neither the evil you identified (the painful public performance) nor the justifying good (the lesson in responsibility) has obtained. This seems to show that the evil in question is, in fact, necessary for the obtaining of the good. (Not metaphysically or nomologically necessary, of course; I trust you mean ‘necessary’ in this counterfactual sense, given your examples.)

    But perhaps you had a different good in mind (this is suggested by some of your phrasing in the quote above). Perhaps the relevant good is the *opportunity* to learn a lesson in responsibility, not *actually* learning a lesson (that’s a further good). The problem now is that permitting the evil of a painful public performance is not necessary (in any sense) for that good to obtain. Johnny’s dad could provide Johnny with the opportunity to learn a lesson (so the good is already achieved) and then prevent Johnny from playing in the recital after the opportunity is squandered (so the evil is prevented). Allowing Johnny to play would make sense only if there is some further good he has in mind (actually learning a lesson in responsibility, perhaps), a good which he judges a painful public performance will be a necessary means.

    In short, I need help finding some good and some evil that are such that the *permission* of the evil is necessary (in the relevant sense, which seems to me captured by counterfactuals) for the good to obtain, but the evil itself is not necessary for the good to obtain. I worry that, for any alleged example, the strategy above will melt it away: a careful look will reveal that the relevant evil obtaining really is necessary for the relevant good, or that the permission of the relevant evil is not at all necessary for the relevant good.

    Can you help me?!

    March 27, 2017 — 9:36
    • Tim Perrine

      Hi Luis,

      Thank you for the comment. As I say in the paper, I don’t have a full theory of God-justifying goods on offer. Your comment pushes me more towards providing one, when I haven’t settled my mind fully. But the distinction is important to my paper, so I’ll try to explain and defend it more.

      First, in the clarinet case I give, I don’t fill in the details a lot. You fill in some details in such a way that there isn’t a relevant evil (e.g. painfully embarrassing public performance). I want to such filling ins that have this result, since it will fail to give us a case to illustrate the distinction we’re interested it. If Johnny takes no pain in horrible clarinet recital, then it seems to me we have a harder time saying we need a justification for it.

      Second, you are correct with the goods and evil I intend to identify in the case. The relevant good G is: learning responsibility (or coming to learn responsibility). The relevant evil E: having a painful and embarrassingly bad public performance, where I assume the child prefers to avoid pain and embarrassing situations.

      I claim two things: the parent’s permission of the evil E is necessary for the good G, but the evil is not. You worry about my defense of this, specifically, that I may be appealing to some counterfactuals in my defense about the permission of E that will equally well apply to E itself, contrary to my intentions. In response, I don’t have anything too fancy in mind. Here’s all I have in mind: G obtains only if the child is given a chance to freely, non-coercively practice. Since the parent’s permission of E requires a chance to freely, non-coercively practice, G requires permission. (We can talk about that if you want.) Now consider: G obtains only if E obtains. That claim is clearly false. The child can develop responsibility and put on a great recital, thereby avoiding a painfully embarrassing public performance. Thus, E is not necessary for G, but the permission of E is.

      In response you might point out the following counterfactual: if the child not experienced a painful performance, then the child would not have learned responsibility for his action. But I don’t think this counterfactual is true. There are two ways that the child could have experience a painful performance: being force to practice by the parent or learning responsibility when given the chance to freely practice. At best, what is true are some ‘may’ counterfactuals: if the child had not experience a painful performance, then the child might have learned responsibility for his actions, he may not have. Further, even if the counterfactual is true, that would be insufficient to refute my claim that the evil is not necessary for the good. Its true that if I kick my dog, she will yell in pain. But that does not show that kicking my dog is necessary for her to yell in pain.

      This may raise the question of what kind of ‘necessity’ I have in mind. That’s a good question. I have two things to say. First, we can introduce a notion of counterfactual necessity like: ‘P is C-necessary for Q IFF Q counterfactually implies P.’ But I think the notion of C-necessary is uninteresting. Even proponents of the counterfactual account of causality wouldn’t accept it or (more cautiously) would find it uninteresting. Second, I guess I have in mind something closer to causal necessity. But causal necessity given all sorts of background information like (e.g.) one cannot take a pill and thereby become great at clarinets. Many authors may want to beef up the necessity from causal to something more when we move from goods that justify parents to goods that justify God. I don’t have anything to say about that at this time.

      Does that help clarify what I have in mind?

      Best,
      Tim

      March 27, 2017 — 11:24
      • Thanks for the reply!

        But I still don’t quite see the real difference the distinction is supposed to track.

        We have settled on the relevant good and evil: (G) learning responsibility, and (E) a painful and embarrassingly bad public performance. Your claim is that the parents’ permission of E is necessary for G, but E itself obtaining is not. My claim is that the parent’s permission of E is not necessary for G either, unless we develop the case into one where E itself obtaining is necessary for G.

        You say: “G obtains only if the child is given a chance to freely, non-coercively practice.” Right. This means that the opportunity for free non-coercive practice is necessary for G, but it does not mean that E is necessary for G, as you grant. However, it also does not mean that *permitting* E is necessary for G. It means that permitting “the opportunity for free non-coercive practice” is necessary for G. Permitting E would be necessary for G only if one of two conditions obtained: (i) E itself was necessary for G, or (ii) permitting E was necessary for permitting the opportunity for free non-coercive practice. But neither is the case (unless we develop the case in ways that makes (i) true, as in my previous comment). You say that “since the parent’s permission of E requires a chance to freely, non-coercively practice, G requires permission”, but this is just false. Permitting E, yes, requires free non-coercive practice; but free non-coercive practice does not require permitting E. So while G requires free non-coercive practice, G does not require permitting E. It just requires free non-coercive practice, unless, of course, things develop in ways where E obtaining becomes necessary for G.

        The key is to see that providing an opportunity for free non-coercive practice is not the same as permitting E, and also that permitting E is not necessary for providing an opportunity for free non-coercive practice. So I still don’t see a case where, for any E and G, the permission of E is necessary for G, but E itself obtaining is not.

        I get headaches when I think about these kinds of things!

        March 27, 2017 — 12:44
        • Tim Perrine

          Hi Luis,

          Thanks for the reply. The distinction I think is intuitive but hard to analyze. Just to fix terms, we’ll let ‘E’ be a painful public performance, ‘G’ be learning responsibility, and ‘F’ be free, non-coercive practice.

          I claim three things. E is not necessary for G: by learning responsibility and practicing, the child can avoid a painful performance. You agree. I also claim G requires F. You agree. The issue, then, is my further claims that G requires the permission of E.

          Let ‘P’ be ‘permitting E.’ Now either (i) P implies E or (i) it does not. I claim G requires P. If P requires E, then I would have to claim G requires E. But clearly G does not require E. Thus, to be consistent, I cannot hold that P implies E. I must instead hold that P does not imply E. And so I will. (That’s of course no argument for someone else to hold that P does not imply E. But I’m trying just to develop the distinction in a coherent way.)

          But if P does not require E, then what is P? I think a natural proposal would be that P is F. Permitting E is the same as allowing free, non-coercive practice. If P = F, then given G requires F, clearly G requires P as I claim.

          You might object that this means that E requires F, so if F = P, then it follows that E requires P. But that just amounts to the claim that the evil requires the permission of the evil, which doesn’t strike me as problematic.

          I can hear the following objection, though: it seems strange to say that one could “permit E” without “E” actually obtaining. As a linguistic matter, I’m not sure; the impression otherwise might just be because we’re using short notation. For instance, I think the parent in these situations could meaningful say things like “I permitted that he fail” or “I allowed that his recital be horrible” but I think they could say in the same breath things like “I permitted that he fail, I’m just glad he didn’t” or “I allowed that his recital be horrible, by not forcing his to practice, but boy am I glad he did practice!”

          But more importantly as a philosophical matter I think it is appropriate to let “permitting E” stand for “allowing for the various necessary conditions for E” or “the situation in which E might have occurred” or something like that. Much evil (but not all) evil is caused by free human action. In that way, God does not actively bring about that evil. Rather, God permits certain situations where that evil could happen. It is natural to ask: why does God permit those situations where that evil could happen? After all, the thought goes, God could keep those situations from happening, thereby keeping the evils from happening. But the focus here is on God’s permission of situations where the evil could occur. Some such situations are also consistent with the evil not occurring.

          Best,
          Tim

          March 28, 2017 — 11:10
          • Thanks for the engaging discussion, Tim. I think we are making progress!

            Some terms:

            E: painful public performance
            G: learning responsibility
            F: freedom to practice or not

            We agree that E is not necessary for G and that F is necessary for G. The tricky question is whether “permitting” E is necessary for G. It’s clear to me now that understanding exactly what is involved in “S permits E” is of central importance.

            So you suggest the following account: permitting E “is the same as allowing” F. But that seems extremely odd and implausible to me. At best, “allowing F” is the same as “permitting the possibility of E”. Here’s why: I can allow F and prevent E by not letting Johnny perform. It would be very odd to say later, in that case, that I had permitted E, but decided to not allow it to obtain. Permitting E and allowing F seem to be different things to me.

            After thinking about this for a bit, I offer you an alternative account:

            (Permission): S permits “x obtaining at t” if and only if (i) S can prevent “x from obtaining at t”, (ii) S believes that “x obtaining at t” is very likely, and (iii) S does not act so to prevent “x from obtaining at t”.

            This means that it is indeterminate whether or not I have permitted an E until the moment it occurs or would have likely occurred. If I am allowing my son to steer the car, and the car is slowly approaching the curb, and I wait for him to correct the wheel, and wait, and wait, and correct it for him at the last second, I have not thereby permitted E. If there is no sign that he is about to crash and at the last second he has a spasm and steers us into the curb, I have not thereby permitted E. If I try to steer the car but my hands get caught in the seatbelt, I have not thereby permitted E. If I notice the curb approaching, and wait, and wait and, at the last second, do nothing only to be surprised by my son steering us away from it, I have, in fact, permitted E even though E did not obtain. All of this seems right to me. (Does it seem right to you?)

            But when we apply this to our case, we do not find an important role for “S permitting E” to play. Since I want G for Johnny, I allow F. It is indeterminate, as of now, whether I have permitted E. If Johnny practices and is ready for his recital and I know it, then *whether or not E obtains* I have not permitted it. So this is not a case where permitting E was necessary for G. If Johnny does not practice and is not ready and I know it, then I have a choice to permit or not permit E. But notice that I am now in a context where I can get G for Johnny only if E obtains! Johnny squandered his opportunity to learn responsibility without E. If anyone—me or someone else—prevents E, then there is no G. So here permitting E is necessary for G, alright, but only because E obtaining is necessary for G and E obtains only if I permit it.

            So my worry remains: if God is justified in permitting an evil, then it is because events have evolved in ways such that this evil obtaining is required for some good. If this evil obtaining is not itself required for some good, then God can allow for everything else until the moment of that evil, and then prevent it. I can allow my son to steer the car close to the curb since allowing him the opportunity yo correct the car himself is necessary a good, but I can intervene at the last minute if the crash itself is not necessary. I’ll permit the crash, of course, if it is, in fact, necessary.

            Where am I going wrong?

            March 28, 2017 — 15:52
  • Tim,

    Maybe I’m missing something clear about your view as well. Let’s s see if I can explain my objection better, based on your latest reply:

    You want to hold a position like (iii). So, there is at least one argument from evil (say, AFE1) that you want to reject on the basis of skeptical theism (+epistemic principles, etc.).
    Let E be the evils we observe and that are used in AFE1 (I won’t put them in the background, to make it more clear in light of one of your previous replies).
    AFE1 would say something like: P(G|E&R)<0.05. (G is "God exists")
    So, if you're going to reject it based on skeptical theism but without giving low probability to R, it seems to me you want to reject it by being agnostic about R. Here's the problem I see (given that the evils are in fact accepted; you don't want to reject AFE1 by claiming the evils do not occur).

    Let's assume then P(G|E&R)<0.05 (else, you would reject the argument on grounds not having to do with skeptical theism).

    P(G&R|E)=P(G|E&R)P(R|E)<0.05

    So, P(G&R|E)<0.05

    We have then P(G|E)=P(G&R|E)+P(G&¬R|E)<0.05+P(G&¬R|E)P(G|E)-0.05.

    If you hold that P(G|E)>.9 (for example), that imples P(¬R|E)>0.85.
    Now, you might say it’s okay for you to say that P(G|E) is lower than 0.9 (much lower), and then you raise the probability by means of other arguments. Then it seems to me that you’d be rejecting the argument from evil on grounds that have nothing to do with skeptical theism. In fact, you’d be saying that in the end, the probability of God given E and some other stuff is very high, but that other stuff has nothing to do with R, but with cosmological arguments and the like. An alternative way of saying this is that you actually accept the argument from evil AFE1, and say it lowers the probability of God’s existence considerably, but that is offset by other considerations (such as cosmological arguments, etc.), not related to skeptical theism.

    You might alternatively say you refuse to assign probability to G|E, but I don’t know how that works: You assign very high probability to God’s existence prior to considering E, then you stop assigning probability when you factor in E, and then you come back to assign very high probability when other factors are included? Is that the idea? But that does not seem warranted as far as I can tell. Or do you just say you don’t factor in E at all? But then, why not? And if not, then how do you go about assigning final probability to God’s existence, given that E is part of the info?

    Best,
    Angra

    March 27, 2017 — 14:50
    • Tim Perrine

      Hi Angra,

      Thanks for your patience. I think I finally get the idea (famous last words!). Consider the following equivalence:

      P(G|E) = P(G|R&E)*P(R|E) + P(G|~R&E)*P(~R|E)

      Your thought is this. P(G|E) should get a high probability—like .9—lest one claim that E is strong evidence against theism. But P(G|R&E) should get a very low probability—like .05—lest one claim that skeptical theism is irrelevant to the argument that requires R. But that means the theist who is a skeptical theist needs to make up that .85 probability somehow to be probabilistically coherent. And even in a “best case scenario” where P(R|E) = 1 and P(G|~R&E) = 1, she have to hold that P(~R/E) = .85. But that means in turn that she cannot really be agnostic regarding ~R.

      A couple of thoughts. First, hey! That’s pretty clever. Thanks for pointing it out. Second, I’m not sure what to say about it. I have a couple of knee jerk reactions, but I’m not sure how many are promising. One reaction I have is this. The probability calculus requires some pretty fine structures. But we usually don’t think reasonable belief is as fine as the probability calculus requires. Thus, one might recognize that the probability calculus has certain equivalences but we cannot use them to force one to hold certain attitudes. This amounts to rejecting probabilism, the view that we are reasonable/rational only if we are in line with the probability calculus. I think all of that is probably correct and we should reject probabilism. But even if we reject such a position in general, it is not clear such a reaction will help with this particular situation. In any case, I’ll have to think more about this.

      Finally, a big picture comment. Lots of people—both philosophers and laypeople—think the problem of evil provides really strong evidence against theism, aka evidence strong enough on its own to make it reasonable to believe atheism. Many philosophers I’ve encounter go further and think it is “knock-out” evidence: evidence so strong it “knocks out” theism from serious theoretical contention. But suppose skeptical theism keeps this from happening. Given it, perhaps the existence and distribution is some evidence against theism, but not “knock out” evidence and not evidence sufficient to justify atheism on its own. Nonetheless, to be reasonable in believing theism, one would need further evidence to counterbalance the evidence evil does provide. I think if all that were true, that would be an important position. Further, I don’t think I’d describe that as a situation in which one was rejecting the problem of evil “on grounds that have nothing to do with skeptical theism.” Skeptical theism would just be one part of defending theism, greatly mitigating the force of certain arguments.

      For what it is worth, I think the kind of response I’m describing happens in scientific inquiry all the time. An example. The retrograde motion of Mars was seen as powerful evidence against geocentrisim—if everything revolves around the earth, why does Mars sometime go the other way? The idea that planets don’t themselves move on circular orbits but rather do epicycles on rotational orbits was proposed. That proposal kept the retrograde motion of Mars from being powerful evidence against geocentrisim. But geocentrisim was still somewhat disconfirmed by the retrograde motion of Mars and proponents of that position sought sources of evidence that would offset such disconfirmation.

      Best,
      Tim

      March 28, 2017 — 10:16
      • Hi, Tim

        Yes, you got it!
        Sorry I wasn’t more clear before.

        Regarding probabilism, that’s an intriguing point.
        If you define probabilism as “the view that we are reasonable/rational only if we are in line with the probability calculus.”, then as long as you have a common interpretation of what it means to be in line with it, I would not only reject it, but claim that it’s not rational to embrace probabilism, for the following reasons:

        Let PI1 be the statement “The 22,459,157,718,361,392-th digit of pi is 5”.

        It’s not known whether PI1 is true. However, there are algorithms that can decide whether it follows from some axioms, and eventually a computer will figure out whether PI is true.
        Now, let A be the conjunction of the axioms in question.

        I rationally assign a probability >0.99 to A.
        Earlier you said (following the most common interpretation of the probability calculus, I think): “Anytime P entails Q but not conversely P(P) > P(Q).”
        I could use that one to show the problem, but it’s better to use: “Anytime P entails Q, P(P)>=P(Q)”.

        If we go by that, then either P(A)=<P(PI1), or P(A)=<P(¬PI1).
        If P(A)=0.99
        If P(A)=<P(¬PI1); then I ought to assign P(PI1)=P(Q)”. But the others are treated like other propositions whose probability we ought to assess on the basis of the available information.
        Yet, that solution doesn’t help the case of skeptical theism: the argument I gave still works, since we’re not dealing with one such exception, it seems to me.

        Granted, you might argue that there are other exceptions, but the fact is that the skeptical theist would be openly breaking the rules of the propositional calculus, and not by means of excluding some formulas from the information we have. I don’t think this is justified, and I think the burden would be on the skeptical theist.

        With regard to the big picture comment, I generally agree about the type of response you suggest – i.e., that’s a common and reasonable sort of response.

        However, I’m an atheist (with respect to an omnimax person, i.e., a person who is omnipotent, omniscient and morally perfect), and I don’t think arguments for skeptical theism would change that.
        I don’t want to go on a long argument off-topic, so very briefly, my take reasons for thinking that are as follows.

        1. I don’t agree with the premises of skeptical theism, and I sort of agree with the views of those many philosophers you mention who consider the evidence from evil so strong it “knocks out” theism from serious theoretical contention (no offense :), but it’s my view on this)…but I say “sort of”, because I only think the evidence from evil would knock out theism from my epistemic perspective if it hadn’t been previously knocked-out so to speak, which leads me to
        2. I think the prior of an omnimax person is so almost zero under the assumption that there is an omnipotent, omniscient person (btw, this assumption sidesteps fine-tuning, Kalam, etc., independently of other considerations), so that one would need a lot of evidence to raise theism from almost zero (note: I say “prior”, but of course I’m using a lot of background info; One might alternatively say the prior is higher but then the probability gets lower, depending on what kind of evidence goes into the background. But it’s background from my perspective).

        Now, if one rejects the premises of skeptical theism, the final probability goes even lower.
        If one accepts them, at best the probability of theism gets stuck at the level of the prior, and that’s almost zero already.
        So, in my assessment, it won’t work. But that said, if skeptical theism were accepted, or even if its premises were given a 0.5 probability or higher, some of the arguments against (and probably even for) theism would have be very different from what they usually are, so the matter would still be important in the dialectical context of present-day philosophy of religion.

        Best,
        Angra

        March 28, 2017 — 18:57
  • Leave a Reply to Tim Perrine Cancel reply

    Your email address will not be published. Required fields are marked *