I. The Problem of Gratuitous Evil: What is the Problem?
The problem of evil, if successful, shows that the existence of evil in the world tarnishes the moral perfection of an omnipotent and omniscient being. If an all-powerful and all-knowing being exists, then He would know about the possibility of every evil and could prevent all of them from happening. So, the fact that evil exists shows that He is not morally perfect or caring enough to have prevented the evil from happening. Assuming that such an omnipotent and omniscient being is a morally responsible agent, then He is morally blameworthy for allowing evil to occur, and, therefore, He lacks the moral perfection necessary for being an all-perfect God.
However, can we blame a morally responsible agent simply because she has brought about a bad or evil state of affairs? A theist may grant that God is (directly or indirectly) causally responsible for anything that happens, including all the bad/horrible states of affairs that obtain in the world. However, from the fact that an agent S is causally responsible for an action A (assuming it is a fact), it does not necessarily follow that S is also morally responsible and blameworthy for performing A. For example, if I have a seizure and, as a result, I fall and break a window, while I am a morally responsible agent and causally responsible for the broken window, I am not morally responsible and blameworthy for breaking the window. In other words, a morally responsible agent might have some specific excusing conditions that defeat or undercut attributions of moral responsibility or blame.
In order for an agent to be morally responsible for performing an action A, she must act knowingly and willingly, that is, she should not act out of ignorance, and/or she must not be coerced into performing A.Footnote 1 Let us assume, for the sake of argument, that acting knowingly and willingly is sufficient to turn causal responsibility into moral responsibility. If so, then when it comes to God, it appears that if God is causally responsible for a state of affairs, then He is also morally responsible for it, because it is impossible for an omniscient being to act out of ignorance, and because it is impossible for an omnipotent being to be coerced to act against His will.
However, is it true that if an agent is causally and morally responsible for performing an action that leads to the obtaining of some bad/horrible state of affairs, then she is an appropriate subject for moral blame?
For example, we grant that a parent who knowingly and willingly punishes her child is causally responsible for inflicting pain and suffering on her child, and, therefore, she is also morally responsible for the action. But is she morally blameworthy for doing so? It depends! If she has a morally sufficient reason for proportionately punishing her child, then she is not morally blameworthy for doing so. For example, the parent might rightfully argue that the proportionate punishment was necessary to teach the child good behaviour and ultimately it contributed to his well-being.
Thus, one might argue that if the evil is tied to a good, and the state of affairs that includes both the good and evil is (significantly) better, ceteris paribus, than the one that lacks both, then the evil-part is defeated by the good-whole. In other words, the good-whole provides a moral justification for its evil-part. In such cases, while the agent who knowingly and willingly causes the good-whole is causally and morally responsible for the evil-part too, she is not morally blameworthy for bringing about the evil-part.
But would such a moral justification be available to God? If the relation between the evil-part and the good-whole is contingent — that is, if it is possible to bring about the good-whole without permitting the evil-part — then an omnipotent God, presumably, can bring about the good-whole without permitting the evil-part, and, as a result, the good-whole cannot provide a moral justification for God to permit the existence of the evil-part. But, if the relation between the evil-part and the good-whole is necessary — that is, if the evil-part is a necessary condition for the good-whole — then no one, including an omnipotent God, can bring about the good-whole without permitting the evil-part.Footnote 2 In that case, the good-whole might provide God with a moral justification for permitting the evil-part (Adams, Reference Adams1999, pp. 20–21 and 28–29; Chisholm, Reference Chisholm1968–1969).
So, God would have a morally sufficient reason for permitting the existence of some evil E, if E satisfies the following conditions: either (a) E leads to a greater good (GG) which significantly outweighs E, and E is a necessary condition for obtaining GG; and/or (b) E prevents a graver evil (GE), which is significantly graver than E, and E is a necessary condition for preventing GE. We may say that if E does not satisfy either one of these conditions, then E is a case of gratuitous evil. (Let us call this definition of gratuitous evil D1.Footnote 3) Thus, according to D1, a gratuitous evil is an evil that either (i) is not a necessary condition for any significantly greater good and/or (ii) is not a necessary condition for preventing any significantly graver evil.
However, some philosophers add another qualification to D1, according to which gratuitous evil also does not prevent an evil that is “equally bad” (e.g., Rowe, Reference Rowe1979, p. 336). Let us call this added qualification Q, and the resulted new definition, D2. Therefore, D2 is a conjunction of D1 and Q.
D2 implies D1, that is, if E is a gratuitous evil according to D2, it is a gratuitous evil given D1 as well, but not vice versa.
So long as an omniscient and omnipotent God does not allow the existence of gratuitous evil, His moral perfection is not tarnished. But if one grants that there is at least one case of gratuitous evil in the world, then it appears that God's moral perfection is tarnished, and that, in turn, indicates that an all-perfect God does not exist.
So, the general structure of an Argument from Gratuitous Evil (AGE) against the existence of an all-perfect God is as follows:
1. If an all-perfect God exists, then no gratuitous evil exists in the world.
2. Gratuitous evil exists in the world.
3. Therefore, an all-perfect God does not exist.Footnote 4
My assessment of the AGE relies on the concept of ‘moral tragic dilemmas.’ Based on this concept, I argue for the two following claims:
First, given D1, one may grant that the second premise in the AGE is true, that is, gratuitous evil does exist in the world, however, the first premise is false. There are philosophers who have already questioned the truth of the first premise (e.g., Van Inwagen, Reference Van Inwagen2006, pp. 99–106). But I will question this premise for different reasons.
Second, given D2, one may grant that the first premise is true, that is, an all-perfect God permits no gratuitous evil, however, there are reasons to believe that the second premise is/could be false. Of course, to question the mere existence of gratuitous evil is a common strategy that has been adopted (among others) by sceptical theists (e.g., Wykstra, Reference Wykstra1984). In this article, I will not reiterate their arguments in detail. But I suggest that the concept of ‘moral tragic dilemmas’ can offer an additional reason that there might not be any gratuitous evil in the world.
Before I present my arguments, I need to clarify what I mean by moral dilemmas in general, and tragic dilemmas in particular.
II. God and Moral Dilemmas
(i) What Is a Moral Dilemma?
A moral dilemma, roughly speaking, is a situation in which an agent S morally ought to do A, and morally ought to do B, but cannot do both. In other words, it is impossible for S to bring about a state of affairs that contains the conjunction of A and B.
But why is S unable to fulfill both of her obligations? Sometimes the impossibility is due to some contingent facts of the world, for example, the reason S cannot fulfill both A and B is due to her limited power or knowledge. For example, suppose I promised my friend to meet him in his office at 4:00 pm. On my way to the meeting, I witness an accident, and I am the only person on the scene who can save the life of the accident victim. But, if I save that person's life, I cannot meet my friend on time. Therefore, it seems that I am facing a dilemmatic situation: on one hand, I ought to keep my promise and meet my friend on time, and, on the other hand, I ought to save the victim's life. But I cannot do both. In this case, the impossibility expressed by ‘cannot’ is simply an expression of some contingent features of the world that prevent me from fulfilling both of my obligations. The fact that I cannot fulfill both of my obligations is due to the fact that (among other things) I have limited knowledge (for example, I did not know that I would encounter an accident on my way, otherwise, I would not have made a commitment to my friend to meet him on that day and time), and/or I have limited power (for example, I am not fast enough to both save the victim's life and meet with my friend on time).
But sometimes the impossibility is due to the fact that meeting both obligations is not logically (in a broad sense) possible. The easiest way to conceive of such dilemmatic situations is when B is simply not-doing-A (or it (necessarily) prevents ‘A’ or brings about ‘~A’). In other words, when S faces a situation in which S ought to do A and S ought to not do A, S is in a dilemmatic situation in which it is impossible (in a broadly logical sense) for S to fulfill both of her obligations.
For example, let us assume that I am morally obligated to not unjustly harm anyone. But if I perform A, I will unjustly harm a person X due to my single act of performing A, and if I do not perform A, I will unjustly cause equal harm to another person Y due to my single act of refraining from performing A. In such a situation, it is logically impossible for me to bring about a single state of affairs containing (A & ~A).
Now one may argue that dilemmatic situations are impossible. For example, Kant writes:
A conflict of duties (collisio officiorum s. obligationum) would be a relation between them in which one of them would cancel the other (wholly or in part). – But since duty and obligation are concepts that express the objective practical necessity of certain actions and two rules opposed to each other cannot be necessary at the same time, if it is a duty to act in accordance with one rule, to act in accordance with the opposite rule is not a duty but even contrary to duty; so a collision of duties and obligations is inconceivable (obligationes non colliduntur). (Kant, Reference Kant and Gregor1996, § 224).
Kant's argument assumes that one ought to perform a certain action only if one can do so. In other words, ‘ought’ implies ‘can.’ Let us call this the ‘Voluntarist Principle.’ Assuming this principle, one may argue as follows:
1. If in a dilemmatic situation, S ought to do A (‘A’ being one of the two conflicting obligations), then S can do A. (Given the Voluntarist Principle.)
2. If in a dilemmatic situation, S ought to do B (‘B’ being the other conflicting obligation), then S can do B. (Given the Voluntarist Principle.)
3. If a dilemmatic situation is possible, then S is morally obligated to do both A and B. (Assuming the definition of a dilemmatic situation.)
4. If a dilemmatic situation is possible, then S can do both A and B. (From 1 and 2.)
5. But S cannot do both A and B. (Assuming the definition of a dilemmatic situation.)
6. Therefore, a dilemmatic situation is impossible. (From 4 and 5.)
Some supporters of the possibility of moral dilemmas have gone to great lengths to provide independent reasons to refute the Voluntarist Principle (e.g., Sinnott-Armstrong, Reference Sinnott-Armstrong1988, Chapters 4 and 5; Stocker, Reference Stocker1987). But here, at least for the sake of argument, I simply assume that the principle is true, and instead, I take a different approach.
Kant assumes that ‘ought’ always expresses necessity; in other words, moral obligation is analogous to logical necessity. Let us call it the ‘Strong Sense’ of ought. Given the Strong Sense, in a dilemmatic situation, S must do A no matter what, and S must do B no matter what, and S cannot do both. Understanding ‘ought’ as ‘must,’ in a dilemmatic situation, none of the obligations annuls the other, and the agent faces at least one ‘impossible obligation,’ that is, an obligation that she must but cannot fulfill. The supporters of moral dilemmas take such impossible obligations as counterexamples to the Voluntarist Principle (e.g., Lemmon, Reference Lemmon1962; Trigg, Reference Trigg1971), and the opponents of moral dilemmas take them as reasons to refute the existence of moral dilemmas.
But we have no reason to assume that ‘ought’ always expresses ‘necessity.’ Sometimes ‘ought’ is a prescription that declares an action to be done from the perspective of one among many moral values. Let us call it the ‘Moderate Sense’ of ought. So, when we say S ‘ought’ to do A (in the Moderate Sense), it does not mean that S must do A no matter what. It simply means that, given certain moral values, doing A is prescribed. If so, then we can still have a dilemmatic situation even if ‘ought’ does not express ‘necessity.’ According to the Moderate Sense of ought, a dilemmatic situation is the one in which an agent S ‘ought’ to do A (from the perspective of a moral value V1), and S ‘ought’ to do B (from the perspective of a moral value V2), but S cannot do both A and B (Brink, Reference Brink and Mason1996, pp. 111–114).
Given this analysis, moral dilemmas (in the Moderate Sense) are always resolvable: in a dilemmatic situation of this sort, since it is morally wrong for the agent to fail to do either A or B, she is morally obligated to do either A or B, that is, the disjunction ‘either A or B’ is the agent's must-obligation. In such a situation, sometimes there is a morally relevant feature of the situation that helps the agent to prefer one option over the other; that is, given that moral ground, one option is the morally best choice available to the agent. (Here, to simplify the discussion, I assume that what is morally best is always overall best.) However, we may conceive of situations in which there is no morally relevant feature based on which the agent can choose one option over the other(s). In such situations, from a moral perspective, it does not matter which option the agent chooses so long as she chooses one of them.
However, sometimes in dilemmatic situations, no matter what the agent does, she has done something really bad or horrible. We may call such dilemmatic situations, ‘tragic dilemmas.’ For example, in Bernard Williams’ famous example of Jim the botanist, Jim through no fault of his own is forced to face two options: either kill one innocent person to save the lives of 19 other innocent people, or not kill that one person, and let all 20 innocent people be killed by the captain in charge of an armed group. If Jim chooses the first option, then as a result of his action, one innocent person loses his life, and if he chooses the second option, then due to his decision, 20 innocent people lose their lives. Either way, the outcome is bad/horrible (Williams, Reference Williams, Smart and Williams1973). So, we may say that Jim is facing a tragic dilemma. However, in this case, one may argue that the death of one innocent person, all things considered, may be a morally preferable choice to the death of 20 innocent people and, therefore, it is the morally best choice. If so, then from a moral perspective, these two options are not symmetric, that is, one ‘ought’ (in this case, the first one) overrides the other ‘ought’ (in this case, the second one). We may call such situations in which the agent has a moral ground to choose one option over the other, ‘asymmetric resolvable tragic dilemmas.’
However, we may imagine some tragic dilemmas in which there are two equally compelling ‘oughts’ involved and no morally relevant features of the situation to serve as a ground upon which the agent can prefer one over the other. And since none of them, by hypothesis, can override the other, none of them will turn into a ‘must.’ If the agent performs neither of them, she has done something morally wrong. But if the agent performs either of them (on the basis of some non-moral ground), then she has done nothing morally wrong. In other words, in this case, the disjunction of doing ‘either A or B’ is the agent's must-obligation. We may call such situations in which both options are morally equivalent and the agent has no moral ground to choose among equally compelling ‘oughts,’ ‘symmetric resolvable tragic dilemmas.’
(ii) Is it Possible for God to Face a Tragic Dilemma?
Religious texts suggest that in the process of creating Adam, God found Himself in a dilemmatic situation. According to The Qur'an, on one hand, God acknowledged that His viceroy on Earth (i.e., humans) will do corruption and shed blood, but on the other hand, He acknowledged that they will also carry a precious and unique knowledge that no other creature can possess. Should He create humans and allow corruption and bloodshed on Earth, or should He deprive the world of the unique gift of humanity? (See, e.g., The Qur'an 2/30–31 and 33/72.) The same idea can be found in some religious texts from Judaism's classical period. For example, “Rabbi Berekiah said: When the Holy One, blessed be He, came to create Adam, He saw righteous and wicked [offspring] arising from him. [He said:] ‘If I create him, wicked men will spring from him; if I do not create him, how are the righteous to spring from him?’”Footnote 5
But what does it mean for God to be in a dilemmatic situation (if at all)?
First, we must assume that God is subject to at least some (moral) obligations; that is, there must be certain things He ought to do.
Second, in order for God to be in a dilemmatic situation, it must be the case that He finds Himself in a situation in which He ought to do A and ought to do B, but He cannot do both.
But is it possible for an omnipotent, omniscient, and omnibenevolent God to find Himself in such a situation? More precisely, if God is omnipotent and omniscient, is there any state of affairs containing the conjunction of A and B that He cannot bring about?
In many cases, the fact that an agent S cannot do both A and B is due to some contingent features of the world. For example, as we saw earlier, in the accident example, I cannot fulfill both of my obligations, and it is due to the fact that (among other things) I have limited knowledge and/or limited power. However, no such limitations apply to God: as an omniscient being, He knows everything and presumably foresees the future, and as an omnipotent being, He can bring about any state of affairs that is logically possible.Footnote 6 So, He can either avoid any possible dilemmatic situations, or He can simply dissolve them.
But now consider the following situation: an agent S ought to do A from the perspective of a moral value V1, and S ought to do B from the perspective of another moral value V2, however, doing B would (necessarily) prevent A or bring about ~A. In such a situation, it is logically impossible for S to do both A and B, because to do so requires S to bring about a state of affairs containing (A & ~A), which is logically impossible: As soon as S performs A, she abandons her moral obligation to do B, and if S performs B, she abandons her moral obligation to perform A! For example, in the biblical story of Abraham and his son, Abraham faces a tragic dilemma of this nature: on one hand, as a man of faith and gratitude, he is morally obligated to follow the commands of God as his benefactor master, and sacrifice his son, and on the other hand, as a parent, he is morally obligated to protect the life of his innocent child. The fulfilment of the first obligation logically implies the abandonment of the second one and vice versa. In this situation, it is logically impossible for Abraham to fulfill both moral obligations. We may call such dilemmatic situations, ‘impossible dilemmas.’
However, if it is logically impossible to bring about a state of affairs containing (A & B), then no one, including an omnipotent God, is able to do so. For example, imagine the following scenario: assuming a libertarian conception of free will has either intrinsic value or it is enormously valuable in virtue of the goods that could not obtain without it (all else being equal), a world containing free creatures seems better than a world containing no such creatures.Footnote 7 If so, then an omnibenevolent God ought to create such a world. However (all else being equal), a world with no possibility of evil seems better than a world with such a possibility. If so, then an omnibenevolent God ought to create a world with no possibility of evil. However (assuming a libertarian conception of free will), creating a world with free creatures implies the creation of a world with the possibility of evil. In other words, if God fulfills His first moral obligation (i.e., to create a world containing free creatures), then He will (necessarily) fail to fulfill His second moral obligation (i.e., to create a world with no possibility of evil), and vice versa.
It is not necessarily the case that all dilemmas God may face are due to the possibility of libertarian free creatures. Consider the following possible scenario involving non-human animals (presumably with no libertarian free will) and their sufferings, which are not due to human actions. The scenario is based on the following assumptions: (a) some sufficiently significant good (intrinsic or extrinsic) depends on the existence of some non-human higher-level sentient creatures (e.g., one might argue that the existence of such creatures is a necessary condition for the evolution of human beings); however, (b) (given the laws of nature) it is impossible to create such higher-level non-human sentient creatures without them having the capacity to experience pain and suffering. If so, then God is facing a dilemma: (i) to create a world with no possibility of pain and suffering, or (ii) to create a world with non-human higher-level sentient creatures. If God creates a world in which there is no possibility of pain and suffering — that is, if He creates a world with no non-human higher-level sentient creatures — then some sufficiently significant good will be lost; and if He creates a world with some non-human higher-level sentient creatures, then He has created a world with the possibility of pain and suffering. We can imagine that these two worlds are morally equivalent, that is, they are overall equally good or equally evil. In such a case, it is logically (in a broad sense) impossible for God to fulfill both His moral obligation to obtain (i) and His moral obligation to obtain (ii).Footnote 8
Therefore, it seems clear that God can find Himself in a dilemmatic situation, that is, in a situation in which He ought to do A and He ought to do B, but it is logically (in a broad sense) impossible for Him to do both. Let us call such situations ‘divine dilemmatic situations’ or ‘divine dilemmas.’
So far as I can see, there is nothing conceptually impossible in thinking that some divine dilemmatic situations are ‘symmetric,’ that is, there are two equally compelling alternative oughts, and (by hypothesis) none of the alternatives outweighs the other. In other words, the alternatives are morally equivalent. For example, it is conceivable that both moral obligations have equal deontological force, or the overall balance of good and evil in both cases is more or less equal. Of course, this situation does not have to be due to the agent's lack of knowledge, rather, it could be due to the nature of the situation as we stipulated. In such symmetric divine dilemmas, there is no moral ground, even for God, to prefer one ‘ought’ over the other. And it is conceivable that in some of the symmetric divine dilemmas, no matter what God chooses to do, the outcome is to necessarily bring about a bad/horrible state of affairs, which makes the situation ‘tragic.’Footnote 9 So, a ‘tragic symmetric divine dilemma’ (TSDD) is a conceivable situation in which God is facing two equally compelling moral obligations, A and B, but it is logically (in a broad sense) impossible for Him to do both, and there is no moral ground to prefer either of the options, and whatever course of action He chooses, the outcome is to bring about a bad/horrible state of affairs. The TSDD cases are morally resolvable; that is, God has one must-obligation, that is, He must do ‘either A or B.’ So long as He performs A or B, He has fulfilled His moral obligation; that is, He is not morally blameworthy. However, the tragic aspect of the situation is that, no matter what He does, He has to bring about a bad/horrible state of affairs through no fault of His own.
III. The Problem of Gratuitous Evil: Solution
Assuming D1 (the first definition of gratuitous evil), how should we assess the AGE?
The first premise in the AGE assumes that an all-powerful and all-knowing being who is all-benevolent never allows the existence of a gratuitous evil simply because there is no moral justification for God to do so. Let us call this claim the Main Assumption (MA), and state it as follows:
MA: There is no morally sufficient reason for an all-perfect God to allow the existence of gratuitous evil.
But is MA true?
It may be true that if E (as a necessary condition) leads to a greater good or E (as a necessary condition) prevents a graver evil, then God has a morally sufficient reason to permit E. However, from this statement, it does not follow that the only way God may have a moral justification for an evil E is that E (as a necessary condition) leads to a greater good or prevents a graver evil. In other words, MA is true if the following statement is true:
M: For every x, if x is a morally sufficient reason for E, then E (as a necessary condition) leads to a greater good, or E (as a necessary condition) prevents a graver evil.
But M is false. There are other ways God may have a moral justification for permitting E besides E's leading (as a necessary condition) to a greater good or E's preventing (as a necessary condition) a graver evil. For example, suppose God, through no fault of His own, faces a tragic dilemma in which He ought to perform A and He ought to not perform A, and He must do either A or ~A. Since it is a tragic dilemma, no matter what He does, He is going to bring about something bad/horrible. In such a situation, God is causally and morally responsible for the outcome, but He is not morally blameworthy for doing so. In this situation, it is impossible for God to not bring about evil, and thus, His doing so does not make Him morally blameworthy. Therefore, if we grant that tragic dilemmas are possible, and that it is possible for God to face a tragic dilemma, then it can provide God with an excuse or a morally sufficient reason to permit E, and thus, we have a good reason to believe that MA is false.
We may articulate the argument as follows:
1. There are tragic symmetric divine dilemmas (TSDD) in which, through no fault of His own, God ought to perform a bad/horrible act of A or ought to perform a bad/horrible act of B (when ‘B’ is simply ‘~A’ or it (necessarily) prevents ‘A’ or brings about ‘~A’).
2. In a TSDD, God has a must-obligation to perform either A or B.
3. In a TSDD, God either performs A or B, and, in either case, He has done something bad/horrible.
4. Therefore, in a TSDD, it is impossible for God not to bring about something bad/horrible.
5. If it is impossible for God not to bring about something bad/horrible in a TSDD, then while He is causally and morally responsible for bringing about that bad/horrible state of affairs, He is not morally blameworthy for doing so. In other words, He is excused or morally justified for permitting that bad/horrible state of affairs to happen.
6. Therefore, in a TSDD, God is excused or morally justified (and thus, He is not morally blameworthy) for permitting a bad/horrible state of affairs to happen.
So, even if E is not a necessary condition for a greater good or it is not a necessary condition to prevent a graver evil (i.e., even if E is gratuitous evil, according to D1), God can still be morally excused for permitting such an evil to happen and, thus, He is not morally blameworthy for doing so (and it means that MA is false). Given the possibility of TSDD, the existence of gratuitous evils (as defined here) does not tarnish God's moral perfection, and thus it does not prove that such an all-perfect being does not exist.
Now assuming D2 (the second definition of gratuitous evil), what can we say about the AGE?
Given D2, if God in a TSDD allows a bad/horrible state of affairs to occur, then such state of affairs cannot be counted as a case of gratuitous evil, because according to D2, a gratuitous evil is one that God could have prevented without thereby losing a greater good or permitting an evil that is worse or as bad. Thus, the possibility of God facing a TSDD cannot refute the first premise in the AGE. However, the possibility of God facing a TSDD may offer a reason that the second premise in the AGE might be false.
As a general strategy, the advocates of the second premise offer some cases of evil that appear to be gratuitous, and based on this observation, they inductively infer that it is reasonable to believe that such evils do actually exist. So, the general and simplified structure of the argument for the second premise is as follows:
a) It appears that E (a specific case of evil) is pointless or gratuitous.
Therefore,
b) it is reasonable to believe that E is indeed gratuitous.
To advance the argument, the advocates of this argument may argue for the following claim that thus:
C: The probability of (G/AG) is low.
(Where ‘G’ stands for the existence of God, and ‘AG’ stands for the existence of an apparently gratuitous evil.)
Sceptical theists for a variety of reasons have questioned the validity of the move from (a) to (b) and/or the truth of C (e.g., Alston, Reference Alston and Howard-Snyder1966; Plantinga, Reference Plantinga and Howard-Snyder1996; Wykstra, Reference Wykstra1984).
However, I suggest a rather different approach: if we grant that it is possible for God to find Himself in a tragic moral dilemma, then we have a new way of explaining why the move from (a) to (b) is problematic and/or C is not a decisive reason to reject the existence of an all-perfect God. Given my argument in this article, for every bad/horrible state of affairs E, it is always possible to conceive of an equally bad/horrible state of affairs E’, such that it is logically impossible for God to prevent E and E’ both from happening. In other words, it is always conceivable that God — through no fault of His own — finds Himself in a TSDD, and thus He has a must-obligation to perform either E or E’. Therefore, while E may appear to us as gratuitous, it does not follow that it is indeed gratuitous: E might be a horn of a tragic symmetric dilemma that an all-prefect God is facing.
Now, for the sake of argument, assume that what appears to be a case of gratuitous evil is indeed a case of gratuitous evil (that is, the move from (a) to (b) is justified). Thus, instead of C, we will have the following claim:
C’: The probability of (G/GE) is (significantly) low.
(Where ‘GE’ stands for the existence of a gratuitous evil.)
Also assume that assigning a probability to the statements like (G/AG) or (G/GE) is plausible,Footnote 10 and C’ is indeed true. However, C’ would be a game-changing claim only if GE were the only relevant evidence available to us, because there might be some other pieces of evidence that outweigh the negative evidence of the kind that works against the existence of an all-perfect God. If we grant that it is possible for God to face a dilemmatic situation, then we have a new piece of relevant evidence (i.e., the possibility of TSDD) that if added to our background knowledge will change the probability expressed in C’ in favour of G. In other words, we have no reason to assume that the probability of (G/GE+TSDD) is still low or sufficiently low to justify the denial of an all-perfect God. This approach to the AGE may be called a ‘theodicy of divine tragedy.’Footnote 11
Concluding Remark
In sum, if we accept that it is possible for God (through no fault of His own) to find Himself in a tragic symmetric dilemmatic situation, then it is possible for God to find Himself facing two bad/horrible options while He has a must-obligation to bring about one of them. In such a situation, God cannot help but bring about a bad/horrible evil. In such a situation, while God is causally and morally responsible for bringing about that bad/horrible evil, He has a morally sufficient reason for doing so (i.e., He cannot help but do so); therefore, He is not morally blameworthy, and the existence of such bad/horrible evil does not tarnish His moral perfection. His existence is thus consistent with the existence of such bad/horrible evils — gratuitous or not.
