Meggan Payne and Keith Sims

University of Miami

Western Michigan University

The Problem of Evil and the Alleged Incoherence of Theism

Quentin Smith’s book Ethical and Religious Thought in Analytic Philosophy of Language¹ is a survey of four distinct movements in analytic philosophy of language and a discussion of what various religious and ethical implications those movements have. The conclusion he reaches is that, given the most contemporary of the movements, which he calls ‘linguistic essentialism,’ there is both ethical and religious meaning to life. These meanings are not necessarily like the meanings popularly held, however, especially in the religious case, the popular belief being represented by Alvin Plantinga’s theistic argument. Smith argues in his book both that Plantinga’s free will defense against the argument from evil is inconsistent with the linguistic essentialism it also requires, and that there is a sound argument from evil, which conclusively proves that there is no monotheistic God as traditionally conceived. In this paper, we will argue against both of these claims; we will endeavor to show that Plantinga’s argument is not inconsistent with linguistic essentialism, and that Smith’s argument from evil is not truly sound.

Linguistic essentialism, as characterized by Smith, is the movement in philosophy of language which is concerned with different sorts of essences or properties, as well as with modal logic, the S5 system in particular.

Important concepts about essences are nontrivial and trivial essences on one hand and logically necessary properties on the other. To make sense of theses concepts, another concept, that of rigid designation, is required. Logically necessary properties are properties without which a thing would not be the sort of thing that obeys laws of logic. They are, in other words, properties that all things have. Self-identity is the example Smith gives. Essential attributes are properties that characterize the essences of individuals; they are properties such that if an individual does not have the property, he is not the same sort of thing. For example, an essential attribute of a person is a ‘being with rationality,’ because if something is not rational (or at least potentially rational – we don’t want to pick any fights on this count), then it is not a person, though an individual might require properties in addition to ‘a being with rationality’ to be a person. Trivial essences are attributes that contain a reference to the individuals that the properties are predicated of; for example, ‘being identical to Quentin Smith’ is a trivial essence of Quentin Smith. Note that this reduces, for Quentin Smith, to the logically necessary property ‘being self-identical.’

Rigid designators are required to make sense of these other terms because they require that properties be true of individuals across different possible worlds. For example, suppose that there is a person with the essential attribute of being ‘a being with rationality.’ What it means to say that if this being is not rational, then it is not a person, is that in every possible world where that being is a person, it is a rational being. To be able to say this, one must be able to talk about identity across worlds, which is what a theory of rigid designation provides. Rigid designation is the thesis that names do not vary their reference from world to world.

We have given a very brief description of the basic tenets of the movement in philosophy of language referred to by Smith as linguistic essentialism. This brief discussion will suffice here, as our purpose is not to argue for or against linguistic essentialism, but to discuss Smith’s argument from evil and Plantinga’s free will defense, both of which assume and require linguistic essentialism. Hence, we now leave this discussion, and turn to Plantinga’s free will defense. We will assume linguistic essentialism for the rest of the paper.

Plantinga’s free will defense is a response to the logical problem of evil, which is the claim that the two sentences

G: God exists and is wholly good, omnipotent, and omniscient

E: There is evil

are inconsistent. Plantinga responds to that claim by making the counterclaim that there is a third sentence, consistent with sentence G, that together with sentence G entails sentence E. His candidate for that third sentence is

p need not be actually true, as Smith rightly claims. Plantinga’s defense will work as long as p is possibly true, given that worlds containing free agents are better than worlds containing only perfectly-acting automatons.

The implication of God’s being as described in sentence G is that the world he actually created must be the world in which he did all he could do to make the world the best place it is. The person offering the argument from evil claims that the world could have been better had God simply made it so that no one could actually do evil; the traditional free will defense has been that a world in which no one is free enough to choose to do evil is worse than a world in which some things are free to choose and do so choose. The response, then, is that it could have been the case that God created a world in which people are free to choose to do evil but as a matter of fact do not so choose, and that that would have been the best possible world. Plantinga’s p is a response to this last claim; it amounts to the claim that it does not make sense to talk about people as a matter of fact always choosing to do right.

He asks his opponent to consider a pair of counterfactual conditional sentences:

CA: If Curley had been offered $20,000, he would have accepted the bribe.

CR: If Curley had been offered $20,000, he would have rejected the bribe.⁴

According to Plantinga, one or the other of these sentences must be true. If God actualizes⁵ a world in which the antecedent is true, Curley is still free to make his choice, and it is possible that he will choose the evil choice. But even on the assumption that he actually did make the evil choice, it was still open to him before the fact to make the good choice. Thus God could not have actualized a world in which Curley both was free and in which he always actually made the right choice.

That transworld depravity is true does not imply, Plantinga argues,⁶ that there is no possible world in which Curley both is free and always makes the right choice. It is just that God could not actualize any of the worlds in which that is the case.

Smith’s criticism of this defense of Plantinga’s is that it both requires and violates linguistic essentialism. To talk about Curley as the same individual in whatever numbers of possible worlds in which we wish to discuss him, we need the name ‘Curley’ to refer to the same individual in whatever world, so we need rigid designators. Likewise, we must have individual essences; we must be able to talk about properties that Curley has without which he would not be Curley. Curley would have these essential properties in whichever world of which he was a part. Most relevantly to our example, perhaps, he is essentially the kind of scum that would take a bribe.

Plantinga’s defense also requires a possible worlds theory of counterfactuals, which was developed within the linguistic essentialism movement. Smith claims that it is on this count that Plantinga goes against the movement.⁷ A possible worlds theory of counterfactual truth posits possible worlds, which are maximal propositions (every proposition or its negation is true in a maximal possible world). For a counterfactual to be true, it must be that in the possible worlds most like the actual world, or in other words, the possible worlds similar to the actual world in all respects except for in regard to the antecedent of the counterfactual, the antecedent and consequent of the conditional are both true.

Since all Plantinga needs to do is show that it is possible that all free beings suffer from transworld depravity, he only needs CA to be possibly true. But Smith argues that he cannot even do that. He writes,

His problem with Plantinga’s theory is, then, that since both of the worlds CA and CR are equidistant from the actual world, given that Curley is offered the bribe in this world, the conditions for truth of a counterfactual cannot even possibly be met.

Our contention is that Smith misrepresents the theories of Stalnaker and Lewis, such that Plantinga’s free will defense still obtains. Both the theory of Lewis and that of Stalnaker allow for the potential truth of CA, and if CA is potentially true, then transworld depravity may be true as well.

David Lewis’s theory of counterfactual truth depends on what he calls the Limit Assumption, which is that

In other words, suppose the actual world and a counterfactual sentence (CA will work just fine). The Limit Assumption simply claims that there is a limit to how close, or similar, a possible-but-not-actual world can be to the actual world; there is no infinite chain of closer-and-closer possible worlds. So there is a closest-possible sphere (or group) of possible worlds where the antecedent of CA is true.

This does not imply, however, that there is and only can be just one closest possible world. Lewis uses the word "sphere" instead of "world" throughout his entire discussion of the Limit Assumption, and even claims explicitly that there are no closest possible worlds. He offers a counterfactual, which I will render, to take the indexical into account¹¹: "At a place on p. 20 of Counterfactuals there appears a line more than an inch long." The line is actually just under an inch in length. Lewis writes:

So on Lewis’s theory, there is no one closest possible world to ours, given a particular counterfactual antecedent, although there are some worlds that are closer than others. So just on Lewis’s theory, Plantinga’s defense will stand; CA and CR can be in the same "sphere" of closeness without both being false.

Stalnaker’s theory does claim that there is always a closest possible world. Lewis writes in explanation,

This is exactly what Smith says Plantinga’s free will defense violates. If this was the only relevant part of Stalnaker’s theory. Smith would be right and we would have to choose between Lewis’s and Stalnaker’s theories whose is the most representative of linguistic essentialism. Fortunately for Plantinga, however, Stalnaker’s theory also incorporates the law of Conditional Excluded Middle:

(f > y ) v ( f > ~y )¹⁴

This will be enough to let Plantinga’s defense escape the clutches of linguistic essentialist theory of possible worlds. Let ‘f ’ be ‘Curley is offered $20,000,’ and let ‘y ’ be ‘Curley would have accepted the bribe.’¹⁵ We get the disjunction of CA and CR, such that the sentence "Either CA or CR" is true. But for the disjunction to be true, at least one of the disjuncts must be true. So they cannot both be false, and Smith is again mistaken. Plantinga’s free will defense does not contradict the linguistic essentialist theory of possible worlds.

Having provided an assessment of Smith’s first critique of Plantinga’s free will defense and finding it to be problematic, we now move into a critical evaluation of Smith’s argument against the existence of God based upon what he calls "the logical argument from evil." This argument purports to show that the existence of gratuitous evil is inconsistent with the existence of a wholly good, omnipotent and omniscient God. We shall argue that Smith’s logical argument from evil does not succeed in showing that there is a necessary contradiction between the existence of a wholly good God and the evil that exists in the world.

Recall the discussion of Plantinga’s free will defense above, which suggests that there is no logical incompatibility between propositions G and E as long as proposition p is added. Smith claims that Plantinga’s answer to the problem of evil is unsuccessful due to some erroneous assumptions made by Plantinga about the nature of freedom, which leads him to overlook some crucial possibilities with respect to the kind of world God could have actualized. Plantinga argues that even though God can create free creatures, he cannot cause or determine them to make the right choices. This is because God’s creatures possess external, internal and logical freedom, and so they are not logically determined always to make the right choice. A creature is externally free iff nothing external to the creature determines that he perform or refrain from any particular act. A creature is internally free iff there are no past physical or psychological states that would determine either that he perform or refrain from performing a particular act. Finally, a creature is logically free iff there is some possible world in which he performs act A and another possible world where he does not perform act A.

According to Smith, Plantinga fails to take into account the possibility that God could create a world containing humans that possess external and internal freedom, yet are logically determined to make the right choices in all possible worlds. After all, God is the best of all possible beings, which means he not only possesses external and internal freedom, but is also logically determined to choose the good in all possible worlds. This being the case, Smith thinks it should be possible for God to create a world (W2) containing people who, like God, are internally determined always to choose what is good and reject what is evil. Such a possibility would lead to a necessary contradiction between G and E.¹⁶

We would argue, however, that Smith’s W2 may not be logically possible, that it may not be feasible, and that Smith is making the potentially faulty assumption that God could not have a reason for actualizing a world with evil.

Smith thinks it is possible for God to create a world containing creatures that, like God, are logically determined to choose good rather than evil. However, this notion is based on a mistaken parallel between God’s logical determinism and the possible logical determinism that his finite creatures could perhaps possess. For the sake of argument, we will grant Smith that God could have created creatures that are externally and internally free and yet logically determined to always do the morally right thing (although we are not convinced that this actually makes sense). Even if God were to create such creatures, this would still not be enough to guarantee that there would never be any evil at all in the world. There would be no moral evil. However there would still be an opportunity for surd evil to occur. God can avoid causing surd evil because he is omniscient as well as logically determined, and so he not only desires to avoid evil, but he knows which actions will actions will cause surd evil and which will not. His non-omniscient creatures, however, do not know all the results of any of their actions, and so may unintentionally cause surd evil, even if they are logically determined. If the logically determined creations did end up avoiding completely the causation of surd evil, although they would still be logically determined, it would be a weird sort of determination. They could not be always choosing good because it is good, for they could not know what is actually right from what is actually wrong. They would either get everything right by chance, or they would be given no chance of error, which would make them externally determined.

One reply to this might be that God could have made his creatures omniscient and logically determined. However, this will not work. God cannot make his creatures omniscient because they are necessarily finite by virtue of their having been created. Created entities are contingent and therefore finite by nature. So, since any created being will be necessarily finite in nature, this would prohibit them from being infinite in knowledge. Hence, they cannot be both logically determined and omniscient.

Furthermore, even if one supposes that it is logically possible for God to create a world with free creatures that always choose the good, it may still be the case that such a world is not feasible based upon contingent facts about the interconnected nature of our universe. It is easy to imagine a world where everything is perfectly connected with everything else such that all sets of events coalesce into a unified and perfect existence devoid of evil, but there is a possibility that such a world would not be feasible for God to actualize. Choices are not made in a vacuum, and the events that follow from choices are all bound together in a state of interdependence. For all we know, the actualization of Smith’s W2 might itself be contingent upon copious sets of circumstances and choices that may not be compossible. For example, the good choices of one individual might interfere with the good choices of another individual, either of which could have an impact on the choice of yet another individual. Every free choice made by a free creature, even if it is good, will have ramifications. It may not be feasible for God to actualize a world containing only and all of those people who only make good decisions, and containing no evil. As long as this non-feasibility is even possible, Smith’s W2 is not certainly actualizable.

Finally, it is important to take note of an assumption that Smith makes when he argues for the possibility of W2. This assumption is that God could not have a reason for allowing evil in the world. Smith’s proposition p`, an alternative to Plantinga’s proposition p, is about the logical possibility of God actualizing a world that is completely devoid of evil. However, even if God could prevent evil by actualizing such a world, it would not necessarily follow that God could have no reason for actualizing a world containing evil. We argue that there is another proposition p* that is consistent with G and that together with G implies E:

*: God has a morally sufficient reason for allowing evil in the world.

One might argue that God allowed evil because he knew that it would ultimately bring about a greater good. We think it is likely that only in a world containing evil would God’s creatures be emotionally moved in such a way as to come to a belief in him. It is important to note that by ‘belief’, we do not mean a mere acknowledgement of God’s existence. Many monotheists would argue that God’s purpose for his creatures revolves not only around an acknowledgement of God’s existence, but also around an emotionally enlivened belief and faith in such a god, and it could be the case that the reality of evil is the most powerful circumstance through which a world of otherwise unmovable creatures would be emotionally moved toward the greater good of freely coming to a belief in God. As long as such a greater good is even possible, it not only resolves the logical incompatibility of G and E, but it also provides a reason for why Smith’s alternative W2 would not be a world that God would want to create.

One might object by arguing that God could have accomplished this greater good by creating people with the conception of evil already built into them? In this way, the actual existence of evil could be avoided while the internal conception of evil would lead a person to the greater good of a belief in God.

However, such an objection overlooks a primary facet of our nature, namely, the power of our emotions to motivate us toward the pursuit of greater understanding. The actuality of evil brings about a knowledge of evil that is personal, and such knowledge can move a person toward the contemplation of God in a way that a mere conception would not have moved them. Perhaps in W2, nobody would ever come to a belief in God, whereas they would come to such a belief in worlds containing actual evil.

In summation, we have argued that Smith’s rejection of the proposition that a monotheistic God as traditionally conceived does not exist is not warranted. We divided our case into two lines of critique. First, we argued that Smith was unsuccessful in showing that Plantinga’s linguistic essentialism is incompatible with Lewis’s and Stalnaker’s theories of counterfactuals, which means that Plantinga’s free will defense is at least potentially still viable. Secondly, we incorporated three lines of argumentation against Smith’s logical problem of evil: (1) the argument that Smith employs a mistaken parallel, between an infinite Creator and his finite creation, (2) the argument that W2 is potentially non-feasible, and (3) the argument that God may have morally sufficient reasons for allowing evil in the world. Based upon the above considerations, we conclude that the theist is still well within his epistemic rights in affirming the logical coherence of a monotheistic conception of God.