Mahmoud Morvarid's Article “The Problem of Evil and the Limits of Human Cognitive Faculties”

“The Problem of Evil and the Limits of Human Cognitive Faculties” was the subject of a lecture by Dr. Mahmoud Morvarid, a faculty member of the Analytic Philosophy group at the Institute for Research in Fundamental Sciences (IPM). The session was held on February 19 at the University of Religions and Denominations, organized by the Iranian Association for the Study of Religions and in cooperation with the university’s Faculty of Philosophy. What follows is a summary of this scholarly meeting; it is hoped that the many limitations in preparing and recording a full and comprehensive report of such specialized sessions have not led to an inaccurate or unintelligible transmission of the issues discussed, and that it meets with the readers’ approval:

One of the main parts of the “problem of evil” is the examination of arguments that have been advanced against the existence of God on the basis of evil. These arguments have various versions, and different ways of responding to them have been proposed. Evidential arguments are a class of arguments from evil that have received more attention today.

Skeptical theism, which has been heavily emphasized in the last two decades, seeks to neutralize evidential arguments by appealing to a series of epistemological ideas and doctrines.

Among theistic philosophers, William Alston, Peter van Inwagen, Richard Swinburne, and Michael Bergmann have addressed this subject. In 2014, Oxford published a book titled Skeptical Theism: New Essays. This book discusses these approaches and their new versions in detail.

Types of Arguments from Evil

As you know, two types of argument from evil have been advanced against God: the logical argument and the evidential argument. In the logical argument, the claim is that some of the facts we know about the evils of the world logically entail the negation of God’s existence.

The evidential argument makes a more modest claim: the facts we know about evil do not logically entail the negation of God; however, they make the probability of God’s non-existence greater than the probability of His existence and justify us in believing that God does not exist, even if they do not give us knowledge of God’s non-existence.

J.L. Mackie’s Argument (1955)

(1) Evil exists (2) A being who is omniscient and wholly good would prevent evil as far as it can (3) There are no limits to what an omnipotent being can do (4) Therefore, a being who is omniscient, wholly good, and omnipotent does not exist.

Many criticisms have been leveled against this argument, and premises (2) and (3) are usually considered incorrect. Premise (2) is false; it should be revised as follows:

*(2) A being who is omniscient and wholly good would prevent evil as far as it can, unless it has a morally justifying reason for permitting evil. A justifying reason for evil E could be one of the following: – Producing a greater good whose emergence is not possible except by permitting evil E or an evil equivalent to or more severe than E. – Preventing a greater evil whose repulsion is not possible except by permitting evil E or an evil equivalent to or more severe than E. We call a state of affairs like R a God-justifying reason for permitting evil E if and only if, should an omniscient, omnipotent, wholly good being exist, R would be a justifying reason for it to bring about E. In this case, we call evil E justified.

Also, premise (3) should be revised as follows:

*(3) There are no limits to what an omnipotent being can do, except for limits that arise from metaphysical necessities.

But the conclusion of the three premises (1), *(2), and *(3) is a conditional proposition:

(5) If some evils are unjustified, then it can be concluded that God does not exist.

If we want to conclude from Mackie’s argument that God does not exist, we must have the antecedent of premise five; that is, the following premise is needed: (6) Some evils are unjustified.

If someone can prove this premise, they have reached the conclusion of God’s non-existence. Therefore, it is incumbent upon the proponent of the logical argument from evil to also prove premise (6).

Theodicy and Defense

One of the issues raised in theological systems is the problem of theodicy and defense. Theological systems put forward hypotheses concerning what God's reasons for permitting evils are. These hypotheses are called "theodicy" or "defense." A distinction is usually drawn between theodicy and defense. The aim of theodicy is to offer a hypothesis stating what God's reasons for permitting evils actually are; whereas the aim of defense is merely to state possible reasons for the permission of evils. The most famous theodicies (or defenses) are the following:

a) The free will theodicy; God has granted free will to a number of beings, and the evils of the world are an unavoidable consequence of this free will. Plantinga places great emphasis on this theodicy. b) The soul-making theodicy; God has brought about evils and calamities so that we may attain perfection. This perfection cannot be achieved except through these calamities. c) The regularity and lawfulness theodicy; the existence of these evils is a necessary corollary of regularity and lawfulness in the material world.

Offering a successful theodicy is very difficult. Perhaps the only way to offer a successful theodicy is the scriptural way; that is, after refuting the argument from evil, you present an argument for the existence of God, then appeal to revelation and justify calamities through revelation. In this case, the theodicy will be dependent on the sacred text.

The point here is that to counter the logical argument from evil, neither a successful theodicy nor a successful defense is needed. It is sufficient merely to say that we do not have knowledge of premise (6) (some evils are gratuitous). However, to counter some formulations of evidential arguments (arguments that seek to produce justified belief in the non-existence of God), we may need to offer a successful defense.

The skeptical theist approach is an approach that claims that even if we do not have a successful theodicy or defense, we can still counter the evidential argument, and from this perspective, it is an attractive approach.

Skeptical theism

After the incompleteness of the logical argument from evil became apparent, atheists turned to evidential arguments. There are two ways to respond to evidential arguments: a) Evidential arguments are probabilistic and non-deductive. Although these arguments show that we have evidence against the existence of God, on the other hand, there are also arguments in favor of the existence of God. Now, taking all these arguments into consideration, the total evidence in favor of God's existence prevails.

b) Fundamentally, evidential arguments do not provide any evidence (even inconclusive evidence) that God does not exist. Skeptical theism, in fact, pursues approach (b) in response to evidential arguments. The skepticism referred to in the term "skeptical theism" is not general epistemological skepticism. The skepticism refers solely to the fact that human faculties have limitations in discerning the presence or absence of God's justifying reasons, a limitation that renders us unjustified in accepting the premises of evidential arguments. This approach emphasizes the limits of human cognitive faculties, the result of which is the invalidity of evidential arguments, even if we cannot offer any successful theodicy or defense.

Rowe's Argument

Various versions of evidential arguments have been presented. Just as Mackie's name is linked with the logical argument from evil, the evidential argument is primarily associated with William Rowe. From 1979 to the present, he has presented several versions of the evidential argument, and extensive discussions have taken place surrounding these arguments. William Rowe claims that although we cannot have knowledge of premise (6) (some evils are gratuitous), we can have justified belief. Rowe's argument for this premise is based on examples. He cites the following two instances:

E1: A fawn is trapped in a forest fire and dies after several days; E2: A girl is kidnapped and murdered.

Reflecting on these two examples, we do not discern any justifying reason for permitting these evils. Rowe has shrewdly chosen examples to which the three theodicies are less readily applicable; for instance, in the case of the fawn, the fawn neither possesses free will, nor is its death the result of a free agent's action, nor is it evident that it has a soul capable of development. The only theodicy that might be invoked here is the third, namely the order and lawfulness of nature. But was it not possible for a different law to govern nature such that this evil would not arise? Rowe therefore claims that we discern no justifying reason for permitting E1 and E2. And this failure to find is evidence for the absence of a justifying reason. This mode of reasoning is an "argument from lack of finding to non-existence." I emphasize that this argument is not deductive, not logical, but probabilistic. The very fact that we have not found something increases the probability of its non-existence. In the evidential argument, this much is sufficient.

Such arguments are sometimes good and sometimes bad. One must seek an epistemological criterion for judging the validity or invalidity of this type of argument. Extensive epistemological discussions have been offered concerning the conditions under which such arguments are justified. Wykstra (in several articles), along with Bergmann and Howard-Snyder, are among the theists who have proposed conditions for the justification of these arguments. They believe that these conditions are not met in Rowe's argument, and that Rowe's argument is an instance of a bad case of this type of reasoning. We now present three proposals regarding the conditions for the validity of an argument from lack of finding to non-existence.

Wykstra and the CORE Principle

Wykstra proposes the following principle:

CORE: E is strong evidence for H only if the following conditional proposition is true: if H were false, E would very probably not exist. For example, when you hear a sound from behind a door, that sound is evidence that someone is behind the door only if this conditional is true: if no one were behind the door, I would very probably not have heard a sound. The CORE principle appears intuitive. In applying this principle to Rowe's argument, we must consider what our evidence and hypothesis are. (A) Evidence: We do not find a justifying reason for God to permit E1 and E2. (B) Hypothesis: There is no justifying reason for God to permit E1 and E2.

According to CORE, A can be evidence for B only if conditional proposition C is true: (C) If God's justifying reasons existed, we would very probably discern them.

Wykstra believes that C is false. And because it is false, one cannot move from A to B. The falsity of proposition (C) is explained through the parent analogy. Consider a five-year-old child who understands concepts of good and bad. Her father imposes an evil upon the child. Since the child's cognitive faculties are weak, even if the father has justifying reasons for causing the evil, the child cannot discern them. Therefore, the fact that the child cannot see a justifying reason is not evidence for the absence of a reason. Wykstra argues that the distance between our minds and God's is far greater than the distance between that young child's mind and her father's; it is a finite versus an infinite distance. Hence, if reasons justifying evil existed, we would very probably not discern those reasons. On this basis, Wykstra's response to Rowe has two parts: the correctness of the CORE principle, and the falsity of proposition C. Various claims have been raised in response to Wykstra. Some have said that proposition C is not, in fact, false. Others have said that CORE is fundamentally incorrect. I also believe that CORE is incorrect and that counterexamples can be offered to refute it.

For example, suppose you have entered a lottery. The evidence at your disposal is as follows: one million people have entered the lottery, there is no fraud in this lottery, and only one person will win. It seems that, given this evidence, you are justified in believing that you will not win. But according to CORE, you are justified in holding that belief only if this conditional proposition is true: if you were to win, that evidence would not exist. Yet it is clear that this conditional proposition is false. Whether you won or not, that evidence would exist. Therefore, in this example, your belief is justified, yet the conditional expressed by CORE is not satisfied.

Other examples can also be offered to refute the CORE principle. In light of these criticisms, Wykstra and Perrine put forward another proposal in 2012 and, based on it, criticized Rowe's argument.

Wykstra and the P-CORE Principle

P-CORE: E is strong evidence for H only if the conditional probability of not-E given not-H is greater than 0.5; that is: P(not-E/not-H)>0.5 For example, hearing a sound is evidence that someone is behind the door only if the probability of not hearing a sound, assuming no one is behind the door, is greater than 50%.

How does the P-CORE principle apply to Rowe's argument? (A) We do not find a justifying reason for God to permit E1 and E2; (B) There is no justifying reason for God to permit E1 and E2. According to (P-CORE), A can be evidence for B only if the probability of seeing the reason, assuming the reason exists, is greater than 50%.

Wykstra and Perrine then argue, based on the same parent analogy, that the probability of discerning God's justifying reason, assuming such a reason exists, is not greater than 50%. It seems that P-CORE is also mistaken. A very simple probabilistic example can show that P-CORE is mistaken. Suppose we have two rooms and a cat. A person decides to flip a coin and, based on the coin's result, place the cat either in Room 1 or Room 2. Furthermore, they decide to flip another coin and, based on the result, make the cat either invisible or visible. We have no other information. We open the door to Room 1 and see no cat. The evidence is that I did not see a cat. The hypothesis is that there is no cat in Room 1. How much does this non-sighting make the cat's absence probable? A simple calculation shows that after not seeing a cat in Room 1, the probability of there being no cat in that room reaches 66%; but Wykstra and Perrine's condition is that the probability of seeing the cat, assuming the cat is there, must be greater than 50%. Yet here, the probability of seeing the cat, assuming it is there, is exactly 50%, not higher. So, a case has been found where A can justify B, yet the probability of NOT-A given NOT-B is not greater than 50%. Therefore, the P-CORE principle is also mistaken. Another proposal is Bergmann's, which we do not have the opportunity to present.

The Author's Response

Here we are faced with an argument that seeks to infer the non-existence of a justifying reason for evil from the non-sighting of such a reason. It has been claimed that there are epistemological principles that can regulate when this inference can be made and when it cannot. It seems that all three proposed principles have counterexamples. I will now try to show, through an example, what principle is actually at work here. I have tried to make this example similar to Rowe's argument. Then we must examine whether, in this example, non-sighting can be evidence for non-existence. We will show in two ways that in this example, non-sighting cannot be evidence for non-existence.

Suppose I am faced with a room about which I have no statistical or probabilistic information regarding whether there is a cat inside. Also, unlike the previous example, I have no information about whether the cat is visible at all. I open the door and see no cat.

(A) Evidence: I do not see a cat in the room. (B) Hypothesis: There is no cat in the room.

Two questions arise:

1. Does A raise the probability of B, even slightly? 2. Does A justify me in believing B?

In answer to the first question, it must be said: In general, evidence A raises the probability of B only if the probability of B given A is greater than the prior probability of B; that is, P(B/A&k)>P(B/k)

On the other hand, we have:

Since P(A/k) is less than 1, P(B/A&k) is greater than P(B/k). Therefore, (A) increases the probability of (B).

اما پاسخ به پرسش دوم منفی است. این نکته را به دو شیوه می‌توان نشان داد. شیوه نخست بر این نکته مبتنی است که گزارۀ B منطقاً معادل است با عطف گزارۀ C و .D (C) هیچ گربۀ مرئی در اتاق نیست (D) هیچ گربۀ نادیدنی در اتاق نیست بنابراین در صورتی به باور به B موجه هستم که در باور به D موجه باشم. سؤال این است که شاهد من که ندیدن گربه است آیا می‌تواند مرا در باور به D-که هیچ گربۀ نادیدنی در اتاق نیست- موجه کند؟ مسلم است که نمی‌تواند. اینکه من گربه‌ای را ندیدم مرا موجه نمی‌کند که گربۀ ندیدنی در اتاق نیست. بنابراین شاهد که B باشد مرا در باور به D موجه نکرده و در A هم موجه نخواهد کرد. پس پاسخ ما به پرسش دوم منفی است. شیوه دوم این است احتمال پسینی وجود گربه در اتاق، مطابق معادله زیر محاسبه می‌شود:

از سوی دیگر، طبق فرض مسئله، ما نسبت به احتمال پیشینی وجود گربه در اتاق (یعنی P(B/k))، و نیز احتمال دیدن گربه در فرض وجود آن در اتاق (یعنی P(not-A/not-B &k)) در جهل مطلق هستیم. از آنجا که اصل عدم تفاوت نادرست است، در مواردی که دچار جهل مطلق هستین هیچ احتمال متعینی را نمی‌توان نسبت داد. از این رو، احتمال پسینی نبود گربه نیز نامتعین خواهد بود.

از سوی دیگر، اصل زیر شهودی است:

(PCJ) تنها در صورتی S در باور به p موجه است که احتمال p به فرض همه شواهدی که S دارد بیش از ۰.۵ باشد. بنابراین من نمی‌توانم باور موجه داشته باشم که گربه‌ای در اتاق نیست. استدلال Rowe متناظر با مثال گربه نامتعین است.

(A) ما هیچ دلیل موجه خداوند برای تجویز E۱ و E۲ نمی‌یابیم. (B) هیچ دلیل موجه خداوند برای تجویز E۱ و E۲ وجود ندارد. ما در جهل مطلق نسبت به این هستیم که چه اندازه محتمل است دلایل موجه خداوند برای تجویز E۱ و E۲ وجود داشته باشد. بنابراین P(B/k) نامتعین است. همچنین، در جهل مطلق هستیم نسبت اینکه در فرض وجود چنین دلایلی، چقدر محتمل است آنها را دریابیم. بنابراین P(B/A&k) نیز نامتعین است. پس بر اساس (PCJ) نیافتن دلیل موجه خداوند ما را در باور به (B) موجه نمی‌کند.