Questioning the Aprobability of van Inwagen s Defense

1 Questioning the Aprobability of van Inwagen s Defense Abstract: Peter van Inwagen s 1991 piece The Problem of Evil, the Problem of Air, and the Problem of Silence is one of the seminal articles of the so called skeptical theist movement in response to arguments from evil against the existence of God. However, it is not clear either how exactly the argument in Sections I and II of the article is supposed to go or whether or not there are strong objections which can be levied against the argument found in those sections. In this paper I offer, what I take to be, the strongest presentation of the argument for van Inwagen s position to be found in Sections I and II (which differs notably from Paul Draper s (1996) interpretation of van Inwagen s argument), and argue that even given this more favorable interpretation of van Inwagen s argument, the argument still fails to avoid a strong objection. It seems to me that there is a puzzle present in Sections I and II of Peter van Inwagen s The Problem of Evil, the Problem of Air, and the Problem of Silence and that this same puzzle occurs again in Sections I and II of van Inwagen s Reflections on the Chapters by Draper, Russell, and Gale. One may become aware of this puzzle when asked to explain how exactly van Inwagen s argument for his skeptical conclusion is supposed to go, and in the case of Reflections one may discover a further puzzle if asked why it is that van Inwagen believes that his argument passes by the criticisms raised by Paul Draper in The Skeptical Theist relatively unscathed. 1 In this paper I aim to try and solve this puzzle, concluding that a careful examination of this puzzle reveals that there is no way in which van Inwagen s argument can be structured such that it is both valid and safe from strong objections. 2 1 This belief is perhaps most clearly displayed in van Inwagen (1996), 228-229. Note: references to van Inwagen s The Problem of Evil are cited as van Inwagen (1991), although the pagination indicates pages in Howard-Snyder (1996) and references to Reflections are cited as van Inwagen (1996) 2 In a way my aim here could be seen as the same as the one found in The Skeptical Theist so in order to make clear why this particular project is a valuable addition to what Draper has already done, it seems worth pointing out what I take to be some key differences between what I am doing here and what Draper does in The Skeptical Theist. First, in a certain sense, my argument is more parsimonious because I aim to show that van Inwagen s defense fails on grounds of goodness and aprobability (a term which will be defined shortly) alone, as opposed to Drapers defense which appeals to goodness, aprobability and undefeatability. This is not meant to serve as a reflection of what I think of Draper s criterion of undefeatability. Rather, I choose to argue this way because if I succeed, I will have responded to van Inwagen using only two, seemingly uncontroversial, criteria to do so (the arguments I give in the paper I hope show that satisfaction of the criterion of aprobability is common ground for van Inwagen, Draper and myself). Second, while I think in The Skeptical Theist Draper interprets van Inwagen in the way that the text of The Problem of Evil lends itself most naturally to being interpreted, I do not think this

2 I begin by offering what I take to be the clearest presentation van Inwagen gives for his argument. He writes in Reflections about the argument found in The Problem of Evil that, In chapter 8, I posed a certain hypothesis I shall call D (for defense ). I argued that a very high proportion of the theism & D-worlds are S-worlds (all of them, as far as I can judge) and that no one is in an epistemic position to offer any answer to the question, What proportion of the theism-worlds are D-worlds? If I am right about D, it follows (by the above epistemic principle) that no one is in a position to judge that only a small proportion of the theism-worlds are S-worlds. And, therefore, no one is in a position to judge that the proportion of the HI-worlds that are S-worlds is much greater than the proportion of the theism-worlds that are S-worlds. 3 This quote reflects the locutions that come from treating epistemic probabilities as epistemic judgments about alethic probabilities, but putting momentarily aside any specific conception of epistemic probability, the argument seems to run as follows: 1. The epistemic probability of S 4 on the assumption that theism and D are true is very high (perhaps 1), i.e., P (S/D&T) is very high (perhaps 1) 2. No one is in an epistemic position to assess the probability of D on the assumption that theism is true, i.e., P (D/T) is unknowable 5 3. No one is in an epistemic position to judge that S is epistemically improbable on the assumption that theism is true, i.e. regardless of the potential truth of the matter, no one has the epistemic right to judge that P (S/T) is low. [from premises 1 and 2] 4. Therefore, no one is in an epistemic position to judge that the probability of S on the assumption that HI is true is much greater than the probability of S on the assumption that theism is true, i.e. no one has the epistemic right to judge that P (S/HI 6 ) is much greater than P (S/T). [from premise 3] Given the ease with which I have been able to recreate van Inwagen s argument one may wonder what right I have to claim that the question of how the argument is supposed to go is puzzling. interpretation of van Inwagen showcases the strongest version of van Inwagen s argument to be found in The Problem of Evil. If I can show that even the interpretation that gives van Inwagen the best case fails, this also seems to be of added value. Conversely, a good which Draper s paper has which mine does not is that Draper s argument casts a wider net in terms of the number of defenses it responds to. While I think several things I argue for here could be fruitfully applied in response to other defenses, I have done little here to show why this is so. 3 van Inwagen (1996), 228 4 S stands for a proposition that describes in some detail the amount, kinds, and distribution of suffering the suffering not only of human beings, but of all the sentient terrestrial creatures that there are or ever have been (van Inwagen (1991), 153). T stands for theism. 5 Here unknowable not only refers to specific values, but also to a general range of values such as high, low or middling. I will later substitute talk of unknowability with Draper s concept of aprobability. 6 HI is the Hypothesis of Indifference which is the competing hypothesis that Draper claims fares much better than theism in accounting for the data in S.

3 My response to this fair concern is that outlining this argument helps show that the argument contains a truly puzzling piece, namely, premise 2. Upon this observation, the puzzle becomes discovering what argument van Inwagen is referring to when he claims that he has argued that no one is in the epistemic position to offer any answer to the question,what proportion of the theism-worlds are D-worlds? 7 To answer this question we need to return to van Inwagen s The Problem of Evil. The good news is that returning to The Problem of Evil after having examined what van Inwagen says in Reflections about the argument puts us in a better position than we would otherwise be. This is because initially it is easy for it to appear that van Inwagen is arguing in The Problem of Evil that if P (S/D&T) is very high (or 1) and P (D/T) is greater than 0 that this is enough to infer premise 3 from the argument above. 8 And, as Draper has shown, this is manifestly invalid. 9 But given the presentation of the argument provided in Reflections, we now have a reason to look through The Problem of Evil with our attention directed both towards finding textual support that the argument outlined in Reflections is in fact the argument in The Problem of Evil and towards finding an argument for the claim that none of us are in an epistemic position to assess the P (D/T). But before turning to the text of The Problem of Evil it will be useful to specify a few things. First, if P (S/HI) is not much greater than the P (S/D&T) then I will say that D is good. 7 One might think that another puzzle is why van Inwagen seems to assume the validity of this argument. An implication of Draper s claim that a defense needs to satisfy at least a third additional criterion, undefeatability, is that if Draper is right then van Inwagen s argument is invalid. However, as stated earlier, my goal here is to show that van Inwagen s defense fails even to satisfy the criteria of goodness and aprobability (let alone this third criterion). So for the sake of argument I will grant the validity of this argument and focus on showing why premise 2 lacks the proper support. 8 I am torn between claiming that van Inwagen does in fact argue for this and simply waivers between two different arguments in The Problem of Evil and claiming that his argument as outlined in Reflections is merely poorly presented at points in The Problem of Evil but was always the only argument van Inwagen intended to make. Regardless of which I claim, there is ample textual support that would lead one to think van Inwagen is arguing in this way. See, for example van Inwagen (1991), 156. 9 Draper (1996).

4 Second, if we are unable to assign any specific range of values (like high, low, or middling ) to P (D/T), then I will say that D is aprobable. 10 It is worth noting that the first premise of my reconstructed version of van Inwagen s argument attributes goodness to D (and in fact actually attributes something significantly stronger that goodness to D), and that the second premise attributes aprobability to D. Finally, it will be useful to lay out what exactly the defense that van Inwagen offers is. The defense is comprised of the three following propositions: 1) Every possible world that contains higher-level sentient creatures either contains patterns of suffering morally equivalent to those recorded by S, or else is massively irregular 2) Some important intrinsic or extrinsic good depends on the existence of higher-level sentient creatures; this good is of sufficient magnitude that it outweighs the patterns of suffering recorded by S. 3) Being massively irregular is a defect in a world, a defect at least as great as the defect of containing patterns of suffering morally equivalent to those recorded by S. 11 There is no doubt that van Inwagen attempts to argue that each of these three claims (and the defense as a whole) is epistemically possible. I think there may be room to challenge this, but for present purposes I am willing to grant that this is so. But, as was earlier noted, what van Inwagen needs to show is that D is both good and aprobable. I do not wish to contest that D is good. I believe it is. It is, however, worth noting that van Inwagen first suggests in The Problem of Evil that S needs to be entailed by the D and T (i.e., P (S/D&T) be 1) and then later claims in Reflections that as far as he could judge that P (S/D&T) is 1. 12 But it is clearly false that P (S/D&T) is 1 for at least three reasons. First, van Inwagen seems to assume, at least in the context of the current assumptions, that God adheres to 10 Both these terms come from Draper (1996), 180. 11 van Inwagen (1991), 157. Van Inwagen acknowledges in van Inwagen (1996) in footnote 12 that he was not perfectly consistently in whether or not merely these three propositions or these three propositions plus T constituted the defense. I think he is right to think that which way he defined D would not be crucial for his case, and I will go with the convention he decides to stick with in that footnote, namely, that the defense is comprised only of propositions 1-3. 12 van Inwagen (1996), 228.

5 some kind of goodness maximizing principle. I say this because he seems to assume that God will necessarily bring about the important intrinsic or extrinsic good which depends on the existence of higher-order sentient creatures. I find such an assumption plausible, but it does not seem to be entailed by D and T and seems to be an odd assumption to utilize given the modal skepticism that van Inwagen promotes. Second, given D and T (and the tacit assumption that God will choose to create a world which has the important good referenced in 2), it follows only that God will either create a world with patterns of suffering morally equivalent to those recorded by S or create a massively irregular world. And this of course lowers the P (S/D&T) even further. Finally, even if we added yet another assumption that when faced with a choice to either create a massively irregular world or a world that was potentially only equally defective due to the presence of patterns of suffering morally equivalent to those recorded on S that God would choose to create a world with a pattern of suffering morally equivalent to those recorded on S, this only entails that God will chose to create a world with patterns of suffering morally equivalent to those recorded on S, not that he would create a world with S. And if we grant van Inwagen s claims about the many different ways in which patterns of suffering could be morally equivalent, we must assume that the probability of (S/D&T) is actually quite low. However, because van Inwagen could easily change his first premise so that P (S/D&T) needs only to be good and since I am granting that the general form of the argument is valid, I do not think this point is worth any continued emphasis. 13 I will instead focus my time, first, on trying to identify what van Inwagen s argument is for the aprobability of D and, second, on trying to show why this argument fails. 13 Despite being low, the P (S/D&T) does seem to meet the criterion of goodness because the probability of (S/HI) will also be quite low in what seems to be a commensurable kind of way.

6 As a starting point, it is worth noting that, at the very least, van Inwagen does affirm the conclusion of the argument we re looking for (i.e. he affirms that the defense is aprobable). He writes that, Our story comprises propositions (1), (2), and (3). I believe that we have no reason to assign any probability or range of probabilities to this story. 14 However, this single statement is found amidst the lengthy argument given for the mere epistemic possibility of D which we have already established is clearly insufficient on its own to show that D is aprobable. 15 I wish to turn our attention now to where I think a claim for the aprobability of D makes its next appearance. Upon completing his argument for the epistemic possibility of D van Inwagen writes that, I have urged extreme modal and moral skepticism (or, one might say, humility) in matters unrelated to the concerns of everyday life. If such skepticism is accepted, then we have no reason to accept the evidentialist s premise that an omniscient and omnipotent being will be able so to arrange matters that the world contains sentient beings among whom suffering, if it exists at all, is apportioned to desert. More exactly, we have no reason to suppose that an omniscient and omnipotent being could do this without creating a massively irregular world, and, for all we know, the intrinsic and extrinsic disvalue of a massively irregular world is greater than the intrinsic disvalue of vast amounts of animal suffering (which, presumably, are not apportioned according to desert). If these consequences of modal and moral skepticism are accepted, then there is no reason to believe that the probability of S on HI is higher than the probability of S on theism, and the evidential argument from evil cannot get started. 16 It seems here that van Inwagen counts the aprobability of D among the consequences of modal and moral skepticism. 17 With this in mind we now have part of what could be turned into an argument for the aprobability of D, namely, that if we should be extreme modal and moral skeptics concerning matters that strongly depart from matters of everyday life, then we should 14 van Inwagen, 159. He goes on to write that (With the following possible exception: if we have a reason to regard the existence of God as improbable, then we shall have reason to regard the story as improbable). This claim could appear puzzling because it seems that 1-3 are just as likely to be true on T as on ~T (although obviously without T this story would not make a good defense). I think however the best answer to this puzzle is that this is one of the cases discussed in my footnote 11 where van Inwagen includes T as a fourth statement comprising the defense. 15 van Inwagen makes it clear not only that his main goal is to show that 1-3 are epistemically possible at this point but rather he seems to make it clear that it is his only goal at that point. For example he writes in discussing 3 in D that, we are not required to argue at this point that it is plausible to suppose that massive irregularity is a defect in a world We are required to argue only that for all we know this judgment is correct. (van Inwagen (1991), 161). 16 van Inwagen (1991), 163. 17 I understand most of the things listed as consequents of the conditional in the previous quote whose antecedent is If such skepticism is followed to be expressions of the aprobability of D.

7 conclude that D is aprobable. So while we have made progress, we have also pushed the question back to another level how does van Inwagen argue that we should adopt extreme modal and moral skepticism concerning matters than strongly depart from everyday life concerns? Thankfully, this question is easier to answer. The arguments for this modal and moral skepticism are in fact woven right into the argument for the epistemic possibility of D. 18 The argument for the modal skepticism is presented in a somewhat unorthodox fashion when van Inwagen writes that, I have argued for this [skeptical] view of the epistemology of modal statements (that is, of modal statements concerning major departures from actuality) elsewhere, and the reader is referred to those arguments. In fact, the arguments of those papers should be considered a part of the argument of the present paper. 19 Van Inwagen then goes on to cite three papers in which such arguments are given. 20 Personally, I find a portion of the arguments found in these papers to be rather persuasive and even if I didn t, for the sake of argument, I am willing to grant that they are successful. So I am willing to grant that van Inwagen has made a successful argument for his particular brand of modal skepticism, but that says nothing about whether or not van Inwagen has successfully argued for his particular brand of moral skepticism, which is the result of a different argument. Unlike its modal cousin, the argument for van Inwagen s form of moral skepticism is found right in the text. In fact, there seem to be two such arguments, provided rather quickly and in close proximity. The following are my best attempts at reconstructing these arguments. The first runs as follows: 1. God or evolution has provided us with a capacity to successfully make judgments about size when the judgment is related to practical, everyday affairs, but neither God nor evolution (nor anything else, presumably) has provided us with a capacity to successfully make judgments about size when the judgment concerns cosmic-level states of affairs. 18 And I think it is this interwoven-ness that in part makes van Inwagen s argument so difficult to untangle. 19 van Inwagen (1991), 159. 20 The papers cited are van Inwagen (1977), van Inwagen (1979) and Seddon (1972).

8 2. By analogy, we have reason to think that God or evolution has provided us with a capacity to successfully make modal judgments when the judgment is related to practical, everyday affairs, but neither God nor evolution (nor anything else) has provided us with a capacity to successfully make modal judgments when the judgment concerns cosmiclevel states of affairs. 3. Therefore, by analogy, for all we know God or evolution has provided us with a capacity to successfully make moral judgments when the judgment is related to practical, everyday affairs, but neither God nor evolution (nor anything else) has provided us with a capacity to successfully make moral judgments when the judgment concerns cosmic-level states of affairs. 21 This argument is closely followed by a second which can be restated in the following way: 1. Our intuitions about value are either (1) a gift from God, (2) a product of evolution, (3) the result of social inculcation or (4) some combination of (1) (3). 2. If our intuitions about value are either (1) a gift from God, (2) a product of evolution, (3) the result of social inculcation or (4) some combination of (1) (3), then we cannot be sure that our value judgments should be trusted when applied to cosmic matters unrelated to the concerns of everyday life. 3. We cannot be sure that our value judgments should be trusted when applied to cosmic matters unrelated to the concerns of everyday life. 22 Of course, the obvious problem here seems to be that the conclusions of these arguments indicate only that we should be weak fallibilists about our value judgments concerning cosmic-level states of affairs. If van Inwagen were to accept that weak fallibilism in a particular area is sufficient to warrant extreme skepticism about that area then he would have to accept that the vast majority of us (if not all of us) should be extreme skeptics about almost all (if not all) of our 21 This argument is reconstructed from what van Inwagen (1991) says on page 162. One may wonder why he goes for such a mild conclusion. The reason for this is that this argument is actually nested within a larger argument for the conclusion that there is no prima facie case for the thesis that the actual suffering of beasts constitute a graver defect in a world than does massive irregularity (162) in which this mild conclusion (in conjunction with other premises) is all he needs for that additional conclusion. It also should be noted that the arguments given for modal skepticism, particular in van Inwagen (1977), are much better than this argument. If this argument is what modal skepticism depended on, I should not have granted van Inwagen this point either. 22 This argument is also reconstructed from what van Inwagen (1991) says on page 162 and like the first argument it should be noted that the rather mild claim in the conclusion can be explained by the fact that this argument is actually just a sub-argument for the claim mentioned in footnote 21.

9 judgments. 23 Assuming he does not want to do this, van Inwagen cannot claim that a recognition of the potential fallibility of our moral judgments concerning cosmic-level states of affairs is sufficient for us to conclude that we should be extremely skeptical of our moral judgments concerning cosmic-level states of affairs. Up until this point whenever we have encountered what appeared to be a gap or problem in van Inwagen s argument I have either been willing to modify van Inwagen s point in such a way that the problem was no longer present or looked to the text to see what could be used to fill in the gap or explain why the problem was not really present. But at this point there is no more further place in the text of The Problem of Evil to appeal to. I ve sought to make sure that the strongest possible argument that can be pulled from The Problem of Evil is the one that I examine, and having sought to do this to the best of my ability, I ve not found any argument that successfully leads to the conclusion that D is aprobable (much less any argument that successfully shows that P (S/D&T) cannot be compared to the P (S/HI)). But perhaps one who is interested in defending a van Inwagen-style argument might think that, while it is true that van Inwagen s argument does not succeed, with a little tweaking an argument similar to the one presented by van Inwagen could succeed. I imagine there are a wide variety of ways one might set about trying to do this. 24 However, in the interest of space I will briefly discuss only two such ways and provide reasons for why one should be doubtful that either of these modifications would lead to a successful argument. 23 And as van Inwagen points out, this kind of skepticism would remove any reason to bother discussing the Problem of Evil in the first place (van Inwagen (1991), 162). 24 One way of doing this may be to cash out talk of epistemic possibility in terms of epistemic judgments about alethic probabilities as van Inwagen does in Reflections. I don t think this is a reasonable move and furthermore I think an argument could be created for why this would actually end up harming van Inwagen s case instead of helping it. Such an argument (which would have applications for defenses beyond the one van Inwagen offers in The Problem of Evil ) is rooted in the fact that most claims in defenses could be translated into the general form, It is epistemically possible that p is necessarily true. (If we accept that moral truths are necessary truths, as van Inwagen seems to, then all 3 propositions in D are of this form) and that claims of this form do not fare well to being aprobable on a model in which epistemic possibility is understood as epistemic judgment about alethic probability.

10 First, one might try to show that van Inwagen s brand of modal skepticism alone is sufficient to show that D is aprobable. But this clearly will not do. This results from the fact that of the three statements that comprise D, only 1 is a modal claim. Both 2 and 3 are moral claims. 25 So even if we grant that 1 is aprobable, this is compatible with P (2/T) and P (3/T) both being assigned very low probabilities which would be enough for D to have a low probability despite 1 being aprobable. A second, and I think more promising, way in which one could attempt to modify van Inwagen s argument would be to try to strengthen the arguments van Inwagen offers for his moral skepticism. 26 However, I think this route too will run into difficulties. In order to see why it is important to remember that in assessing whether or not D is aprobable we are assessing whether or not we could reasonably assign a probability or range of probabilities to D on the assumption that theism is true. It is important to remember also that van Inwagen accepts not only that we are able to make at least some moral judgments (namely, moral judgments about everyday matters) but also that the truth of theism may be able to provide us with the epistemic justification needed to believe that we can make value judgments in these select cases. Given these facts, it seems that if we have moral intuitions about cosmic-level moral matters which seem relatively similar to our moral intuitions about everyday matters, then we have good reason to think that our intuitions about cosmic-level moral matters are, at the very least, not so non-veridical that we should consider propositions about cosmic-level moral matters aprobable merely because it is a proposition about a cosmic-level moral matter. Matters would be different if we had no such intuitions about cosmic-level moral matters, but if we do have such intuitions it seems like the best explanation for these intuitions, given that God exists and 25 There may be some trivial sense in which almost all claims are modal claims if that phrase is understood in a loose enough way, but such a usage doesn t seem to match the usage used by van Inwagen. 26 Or perhaps provide completely new arguments for the same conclusion.

11 that our moral intuitions in certain other spheres are, presumably, relatively accurate, is that these intuitions about cosmic-level moral matters will also be relatively accurate. Otherwise it seems like God would be engaging in (or at the very least allowing) a very odd kind of deception. Whether or not this provides a good reason to doubt that a successful argument can be crafted for the conclusion that we should be skeptical about our moral judgments concerning cosmic-level moral matters depends on whether or not one believes that we have moral intuitions about cosmic-level moral matters which appear to be relatively similar to our moral intuitions about everyday moral matters. I, for one, have such intuitions, but my guess would be that van Inwagen would claim that he does not. What now? Given our differing intuitions (or lack of intuitions) are we entitled to draw different conclusions about the reasonableness of claiming we should be skeptical of our ability to make cosmic-level moral judgments? Or should we instead see which position the majority of people are in? I m not quite sure what to say about this, aside from the fact that I think van Inwagen s lack of intuitions, if that is his condition, is atypical. But I am hesitant to draw any further conclusions from that. Perhaps there is something defective about those of us who have moral intuitions about cosmic-level matters which seem similar to our moral intuitions about everyday matters. Perhaps. But even if this were true, doesn t this still leave us with the problem of explaining why God would allow such a bizarre, subtle type of deception to occur perhaps especially trouble upon realizing that these defective intuitions may end up playing a role in human arguments for or against God s existence? I think the implications of this potential stalemate will differ depending on one s opinion about who, if anyone, has the burden of proof in discussing arguments from evil. For the time being though I think it is safe to say that van Inwagen s moral skepticism has not been properly motivated.

12 To review, I ve sought here to identify the best version of the argument for skeptical theism present in van Inwagen s The Problem of Evil, the Problem of Air, and the Problem of Suffering and have concluded that there is no version of the argument to be found which is both valid and free from strong objection. I ve also briefly looked at some additional ways in which a defender of a van Inwagen-style argument might try and tweak the argument to make it more successful and have outlined what I take to be the difficulties that one would encounter in trying to do so. So for the time being I remain skeptical about the success of an argument which depends upon certain moral propositions being aprobable, but as we have seen, one s skepticism is no guarantee of one s correctness.

13 Works Cited Draper, Paul. The Skeptical Theist in The Evidential Argument from Evil, ed. Daniel Howard- Snyder. Bloomington, IN: Indiana University Press, 1996, 219-243. Seddon, George. Logical Possibility, Mind 81 (1972): 481-494. Van Inwagen, Peter. Ontological Arguments, Nous 11 (1977): 375-395. Van Inwagen, Peter. Review of The Coherence of Theism by Richard Swinburne, Philosophical Review 87 (1979): 668-672. Van Inwagen, Peter. The Problem of Evil, the Problem of Air, and the Problem of Silence, Philosophical Perspectives 5 (1991): 135-165. Reprinted in The Evidential Argument from Evil, ed. Daniel Howard-Snyder. Bloomington, IN: Indiana University Press, 1996, 151-174. Van Inwagen, Peter. Reflections on the Chapters by Draper, Russell, and Gale in The Evidential Argument from Evil, ed. Daniel Howard-Snyder. Bloomington, IN: Indiana University Press, 1996, 219-243.