| The Evidential Argument From Evil Part 3 |
 |
| The Skeptical Theist By Paul Draper In this article Draper attempts to discredit the arguments created by skeptical theists. Draper claims that these skeptical theists have serious impediments that they must overcome before they can claim that they have solved the evidential problem of evil. Draper argues that the basis of the skeptical theists argument is the �skeptical thesis #1� or �ST1� for short. ST1: Humans are in no position to judge directly that an omnipotent and omniscient being would be unlikely to have a morally sufficient reason to permit the evils we find in the world. This is the argument that William Alston puts forth. He (and others) argues that we are in no position to know if God has sufficient reason to allow evil to occur. This is where Draper first takes issue. He is troubled not by ST1 itself, but the inference made when the skeptical theist goes from ST1 to the conclusion that all probabilistic arguments from evil fail. He compares it to people who attack one teleological, cosmological, or ontological argument and then conclude that THE teleological, etc� fail. In essence, Draper is accusing them of committing the straw man fallacy. He accuses Alston of doing exactly this, by focusing on William Rowe�s argument and then making generalizations about THE argument. Draper argues that this is a large inference to make. However, Draper says that this occurs often in theistic thought and makes reference to the logical argument of evil, from which probabilistic arguments arose. In this argument, let E stand for evil and G for theism. You can disprove G by disproving the following statement M. M: An omnipotent and omniscient being would have a morally sufficient reason for allowing E. However, theists who argue, that given our cognitive limitations we are in no position to prove that M is false, defeat this. ST1 is very similar to this in the fact that it offers an argument against the probabilistic argument form evil. We are tempted to make the same generalization that we made in the logical argument. However, Draper says that if we do then we are overlooking one major point. The logical argument and the probabalistic arguments are not the same in their make-up and therefore arguments against them will not work in the same ways. Many theists also use ST1 in attempts to defeat arguments that appeal to unjustified evils (an example of this would be the argument advanced by William Rowe). Rowe picks an inscrutable evil and argues that the following statement is prima facie probable: Q: No good state of affairs is such that an omnipotent, omniscient being�s obtaining it would justify that being�s permitting E. Theists find themselves in a situation that if they can show ST1 is true, then Rowe�s argument would be defeated. Therefore, the opposite also seems true. In order for a probabilistic argument to work then it would seem that you would need to defeat ST1. However, this is not correct. Draper argues that his own probabilistic argument does not need to defeat ST1. He claims that �Humean probabilistic arguments� could work. These are arguments like Hume�s argument from evil, which states that �Theism is prima facie improbable because the hypothesis that the first causes of the universe are neither benevolent no malevolent explains the pattern of pain and pleasure in the world much better than theism does.� Although Draper believes that Hume�s particular argument fail, he believes that similar arguments can work. Draper says that the theist must no create a new skeptical principle; we�ll call it ST2. ST2: Humans are in no position to compare theism�s ability to explain certain facts about good or evil to some other hypothesis�s ability to explain those facts. Draper spends the rest of his essay determining if ST2 could be used against his particular argument from evil. To begin with, Draper lays out his argument from evil. His central claim follows: C: O is antecedently much more likely on HI than on G Draper says that �O stands for a statement reporting the observations and testimony upon which one�s knowledge about the pattern of pain and pleasure in the world is based�. The term �antecedently� means �independent of the observations and testimony O reports�. HI refers to the now familiar Hypothesis of Indifference. Draper abbreviates C to the following: C: P(O/HI) >! P(O/G) Draper now says that we can modify ST2 to directly apply to his argument. This gives us ST3: ST3: Humans are not in a position to compare the antecedent probability of O on G to the antecedent probability of O on HI. If the theist can show that some part of C is wrong, then ST3 would be correct. Draper says that the theist has several methods of attack. One would be to be skeptical of the relevance of the biological usefulness of pain and pleasure, which is one of the fundamentals of Draper�s argument (see section II of �Pain and Pleasure�). A second way to defend ST3 would be to be skeptical about how much theodicies raise (or don�t raise P(O/G). However, with some disappointment, Draper points out that Inwagen (who in an earlier article proposed an attack against Humean style arguments) doesn�t use either of these methods, and instead sticks to �defenses�. Draper says that defenses are just a sub-class of theodicies. To distinguish between the two, and their respective results, he once again lays out the premises of his argument using the term �theistic stories� in place of theodicies. C1: Independent of the effect of theistic stories on P(O/G), P(O/HI) >! P(O/G) C2: Theistic stories do no significantly raise P(O/G) In Inwagen�s article he argues that C doesn�t follow from C1 and C2. Furthermore the needed premise, C3, is false. C3 being: C3: Theistic stories do not render ST3 true. A theistic story is a �successful Theodicy� if it makes C false by raising P(O/theism) significantly. A �successful defense� justifies skepticism about C by rendering ST3 true. Next we must see what conditions are necessary for a theistic story, D, to be a successful defense. One condition is that we would have to not be able to assign any value to P(D/G). The second condition is that P(O/HI) can�t be much greater than P(O/D&G). If a theistic story satisfies the se requirements then it is labeled a �good� story. A successful theodicy would be a story in which we could judge the value of P(D/G), and this value was high enough to reject the claim that P(O/HI) is greater than P(O/G). Defenses also have on other requirement and that is they must be plausible. The statement, �For all I know, the story is true�� is not sufficient. We must also add ��and if it is true then theism accounts for O just as well (or at least not significantly worse than) HI does.� The Weighted Average Principle, or WAP can measure the effect that a defense has. P(O/G )= (P(S/G) x P(O/S&G)) + (P(~S/G) x P(O/~S&G)) Though WAP can show us the goodness of a story, it can�t show us if this goodness and the defenses aprobability (lack of a definite value) is sufficient to establish ST3. To show that it is not Draper uses several analogies. The first involves the color of a house. Brownism (B): other things being equal, John would much rather live in a brown house than in a yellow one. Yellowism (Y): Other things being equal, John would much rather live in a yellow house than in a brown one. John then builds a house and paints it yellow. E will now stand for a statement reporting his choice of color. With this in mind we can conclude, prior to considering any brownistic stories, the following; P(E/Y) >! P(E/B). Now our Brownist offers a Brownistic story: D: The Yellow paint that John bought was on sale; no other color was on sale; and John would rather live in a yellow house than miss a chance to save money on a paint purchase. This represents a good Brownistic story because P(E/Y) is not much greater than P(E/B&D). We will also assume, for the sake of argument, that it is also aprobable. However, the Yelloswist is not defeated, because he can offer another story that would be just as likely as the Brownistic story. Let�s call this a �counter defense�. A: The yellow paint that John bought was not on sale; there was, however, a sale on brown paint; and John had a very strong desire to save money on his paint purchase. What this shows is that there must be a third condition. That is, a story must not be undermined by a counter defense. Such a story would be called undefeated. However, even if a story should meet all three of these requirements Draper doubts if it would be able to establish ST3. One solution might be to create lots of different stories. This weight of numbers might be sufficient to establish ST3. So let us say for the moment that enough good, aprobabilistic, undefeated stories could prove ST3. We must now ask if any such stories exist. Draper does not believe any has been made to date. He goes on to give several examples of stories that fail to meet on requirement or another. It seems that they are especially susceptible to counter defenses. Draper claims that unless the skeptical theist can come up with a good, aprobable, and undefeated story, criticism of his claim, P(O/HI) >! P(O/G), is unwarranted. Draper believes that he has successfully defended his Humean style argument against both Alston�s skeptical thesis and the defenses proposed by Inwagen. Defenseless By: Bruce Russell Russell tries to show us that theistic arguments are �defenseless� against evidential arguments from evil. To begin with, Russell lays out the basic form of the evidential argument from evil. 1. If God exists, there is no gratuitous evil, that is, evil which God would have no morally sufficient reason to allow. 2. But there is gratuitous evil. 3. So God does not exist. He claims that these arguments are either probabilistic or epistemic in nature and start from the fact that after we make a careful observation we can see no reason fro God to allow the suffering that occurs. Furthermore, Russell says that these arguments move from their beginnings to their ends by either induction or abduction; Abduction being an inference to the best explanation. He claims that there are four basic kinds of evidential arguments from evil. These can be seen in the chart below. Inductive Abductive Start from the fact that we see no point to allowing certain (1) (2) instances, amounts, types, or patterns of suffering. Start from the fact that there are certain instances, amounts, (3) (4) types, or patterns of suffering. Russell believes that William Rowe has offered a good example of a type 1 argument. Russell argues that because the goods we know of do not provide a morally sufficient reason to allow evil, we have reason to believe that no goods provide reason and that evil is unjustified. An example of this would be the �blue crow� example. If we make a careful examination of the worlds crow population and fail to find any blue crows, we are then able to say, with confidence, that there are no blue crows. A �crowist� could argue that there are blue crows that are �beyond our ken� that we would miss in our examination. However, we could only believe in this possibility if there was some reason for us to believe that our sample was not representative. In this case, however, we have full reason to believe that our sample was representative and that there are no blue crows �beyond our ken�. Atheists believe that their Type 1 arguments fit this requirement. A criticism of this style of argument is that there might be goods �beyond our ken: that would not fall under Rowe�s representative sample. If we think that our sample might not be representative then we can�t determine if God has sufficient reason to allow evil. He gives us an example of such a situation. We encounter a trustworthy group of aliens who show us an animal that we have never seen before. We will call it a zagaloop. Now, we are not justified in believing that no zagaloops are red, just because none of the examples we saw were red. This is because we have no idea if the sample we saw was representative. Russell agrees that this is a real issue. However, he feels that upon closer examination we can show that these arguments do work. Before he shows us why, he first discusses Type 2 and 4 arguments. To begin with, Russell says that Type 2 and 4 arguments both rest upon the same principle. P1: S is justified in believing p if and only if p is part of the best explanation of what S has observed. Russell clarifies this by saying that �observed should be construed broadly to include, e.g., testimony or accounts a person might have heard or read.� S is justified in believing p relative to S�s observations and S* (not S) is justified in believing not-p relative to S*�s observations. For example, S and S* could look through a telescope at the same spot at night and see nothing. S* might have information that justifies his belief that the telescope is so weak that he couldn�t see a planet, even if one was there. However, S might have information that justifies his belief that the telescope IS powerful enough to spot a planet, if one was there. So, here we have a situation in which both S and S* have the same perception, but are justified in believing different things. What this means is that �P1 makes justification relevant to the total evidence the subject has.� For this evidence to be any good it must be of an objective nature. The only distinction between Type 2 and Type 4 arguments is what they take the explanation to be. The starting point of Type 2 arguments is our failing to see the point of certain instances, amounts, types, or patterns of suffering. In Type 4 arguments it is suffering itself that is the problem. In a Type 2 argument it is claimed that the best explanation of our failure to see relevant suffering is that there is none. Theists offer a counter example in which they say there is an explanation, but it is beyond our ken. So which of the two explanations is the best? For this we turn again to our friend, Bayes�s Theorem. P(H/e&k) = P(H/k) x P(e/H&k) P(e/k) When we look at Bayes�s Theorem in relation to Type 2 arguments e represents �the statement that even after much careful reflection we see no point to certain instances, amounts, types or patterns of suffering.� K is represents �whatever justified background beliefs we have. Finally, T represents �there is a reason for that suffering that is beyond our ken.� Our variables established, we can now use Bayes�s Theorem to compare H and T. 1) P(H/e&k) = P(H/k) x P9e/H&k) P(e/k) 2) P(T/e&k) = P(T/k) x P(e/T&K) P(e/k) The equations who�s right side is greater will be the better of the two arguments. Russell says that if these equations are correct then the second equation P(e/T&k) would be greater than P(e/H&k). This is because if there is a reason beyond our ken for suffering than there would be a good chance that we wouldn�t see it. Russell thinks that this is a very important point for it shows us that the role of �apparently pointless suffering� has lost its importance in the problem of evil. This is because the theistic hypothesis is at least as good as the atheistic hypothesis. In order to make a case atheists will have to take another look at the two hypotheses and determine which of the two is better based upon out �background knowledge.� To illustrate this point, Russell gives us an example in which a person believes that the Earth was only created 100 years ago. Of course Russell point out that there is a lot of evidence that would seem to contradict this view. Books and newspapers exist that are over 100 years old and they recount events that are even older. We can see deep river valleys that were created by erosion over long periods of time. We also have mountains, which have taken millions of years to uplift. However, this person is determined. They argue that when God created the world 100 years ago he also created all of this evidence that shows the Earth�s age. They continue by arguing that God has some great moral reason for allowing a deception of this magnitude, however, this reason is beyond our ken. If it is really true that there would be a reason beyond our ken, then this person�s theory is just as good as ours. However, few people would actually believe this is true. This is because the background evidence for the ordinary belief (that the Earth was created billions of years ago) is much easier to accept than the idea that God has taken such elaborate steps to deceive us. Russell argues that if we can reject the idea of Earth being only 100 years old because it seems unreasonable, then we should be able to do the same for out Bayes�s Theorem equations. Of course, as with the Blue Crow example, we can only draw this conclusion after conducting a search for reasons to justify suffering. To help clarify this point, Russell asks us to take the opposite stance for a moment and assume that even though we fail to find any weighty moral reasons for God�s allowing evil, specifically let us say the rape and murder of a little girl, we are not justified in thinking there are none. Russell says that if one is willing to accept this stance then one must be willing to accept the following situation. That is, �we are also not justified in believing that some human being who could easily have stopped the heinous crime did something wrong in failing to intervene.� The reason is that the same reason that God has for His inaction may be the same reason that our imaginary person has. Russell concludes his discussion over Type 1 and Type 2 arguments by saying that the burden of proof lies with theists. In order to defeat Type 1 and 2 arguments theists must show atheists that there samples are unrepresentative. In the next section Russell focused on Inwagen�s position that theists may be able to create a �defense�. He begins by once again laying out two formulas. In these formulas, �e equals 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.� Once again, T represents theism and HI represents the Hypothesis of Indifference. For the purposes of this discussion Russell chose to ignore k, which is our background knowledge. Instead, he chose to assume that T is just as likely on k as HI is. The result is the two equations that follow. (1) P(HI/e) = P(HI) x P(e/HI) P(e) (2) P(T/e) = P(T) x P(e/T) P(e) According to Russell, �Inwagen says that theists can try to show P(e/HI) is lower, or that P(e/T) is higher than one might suppose.� This would amount to a Theodicy and would require an auxiliary hypothesis h. The result would be that P(e/T&H) and P(h/T) are both high. However, Inwagen has doubts about our ability to create a successful Theodicy. Instead, Russell says that Inwagen attempts to create a defense in which Inwagen shows that we are in no position to judge P(h/T) and so not be able to judge P(T/e). The key is to find a story, h, that is not unlikely on anything we know. If we could do that then Inwagen believes that we would be in no position to judge P(e/T), and therefore not able to judge that P(HI/e) > P(T/e). If we can find an h that is not unlikely then the P(e/T&H) would be high, and for all anyone knows, P(h/T) would be high. Russell the lays out the story, h, that Inwagen uses in his defense. (1) Every possible world that contains higher-level sentient creatures either contains patterns of suffering morally equivalent to those recorded by S (our e), 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 (e). (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 (e). Russell believes that you cannot argue against premise 1 by claming there are possible regular worlds with different natural laws and lacking natural disasters. This is because to do so you would have to take on Inwagen�s Model Skepticism, which is Inwagen�s �view that we can�t be justified in believing that a world is possible without being able to offer a plausible �design� of that world.� Russell feels that he is not ready to take that on. Instead, he tries a different attack. Russell argues that it is ok for us to believe that God, if He exists, would have a choice between creating two different worlds, this one and one where God intervenes more to reduce suffering, however, not interfere so much as to produce massive irregularities. Russell also feels that this is unaffected by Inwagen�s model skepticism because we can easily design such a world. It is just like the actual world with one more instance of suffering prevented by God�s actions. Russell feels that God would be obligated to bring about such a world: ��because a person is always obligated to prevent more rather than less terrible suffering in realizing such an end, provided that preventing that suffering does not appear to approach some threshold beyond which terrible consequences will ensue.� The threshold of course referring to the point at which God�s preventing more evil would either do no good or actually do more harm than good. Although Inwagen has no response for Russell�s argument William Alston does. Russell lays out Alston�s argument as follows. Alston claims that we are in no position to judge that God would be obligated to reduce suffering, this is because it might be necessary for achieving some good that we can not see. According to Alston, goods fall into one of four categories. Goods we know of Goods we don�t know of (Goods beyond our ken) The good is primarily of (a) (b) benefit to the sufferer. The good is not primarily of (c) (d) benefit to the sufferer. Of these four options Russell feels that (c) is the most interesting. These are goods we know of that benefit someone else. Alston argues that atheists can�t argue that one more case of evil was not needed to realize some outweighing good for others, because to do so we would have to know that it was �among the n-percent of cases most worthy of being miraculously prevented.� Alston believes that to know that we would have to survey all cases, past, present, and future to see where this one particular instance ranked. Russell feels that this is unnecessary and points out that there are cases of human intervention ��where if the number of interventions surpassed a certain level such bad consequences would ensue that no further interventions should take place.� He uses an example of police checkpoints to find drunk drivers. If more than �some indeterminate number of people are stopped they will become so hostile�that more people will be killed by those drivers than would be killed by drunken drivers if fewer were stopped.� However, we can be justified in believing we are below the threshold where the bad starts out weighing the good, even if we don�t know exactly where that threshold is. Russell then applies this to God�s situation. He feels that God, if he exists, ��would reduce by at least one instance the terrible suffering there is, since there is no reason to think that doing so would have awful results and good reason to think it would have very good effect.� Some would ask where the stopping point would be, how many evils would God have to prevent for it to be enough. Russell feels that the stopping point would be a threshold where if more evil was prevented, goods would be lost or greater evil would be allowed. In the conclusion of his passage, Russell takes one last look at Inwagen�s argument. Russell concludes that Inwagen�s defense, h, is improbable, since we can imagine a possible world that has less evil. Therefore, T&h combined is improbable. We can then use Bayes�s Theorem to conclude that P(HI/e&K) > P(T&h/e&K). (1) P(HI/e&K) = P(HI/K) x P(e/HI&K) P(e/K) (2) P(T&h/e&K) = P(T&h/K) x P(e/T&h&K) P(e/K) Therefore, Russell believes he has shown Inwagen�s hypothesis to be false. He also sees little hope for other hypothesis, unless they did not imply that we are in no position to judge whether or not God is obligated to prevent some evil from occurring. Of course, Russell does allow that if someone could come up with a reason for God�s existence that was weightier than the reasons supporting the view that, if God exists, there would be less suffering because He would be obligated to prevent as much as He could; then Russell�s views would not be probable based upon the evidence. However, Russell concludes that this was not what he was arguing against. Instead he says that if no weightier arguments for God�s existence are presented, then the evidential arguments against God�s existence that he presented should convince both theists and atheists of God�s non-existence. |