A simpler free will defence

Abstract

Otte (Philos Phenomenol 78(1):165–177, 2009) and Pruss (Faith Philos 29(4):400–415, 2012) have produced counterexamples to Plantinga’s (The nature of necessity, 1974) famous free will defence against the logical version of the problem of evil. The target of this criticism is the possibility of universal transworld depravity, which is crucial to Plantinga’s defence. In this paper, we argue that there is a simpler and more plausible free will defence that does not require the possibility of universal transworld depravity or the truth of counterfactuals of creaturely freedom. We assume only (a) that libertarianism is possibly true and (b) that God’s existence is consistent with the existence of free agents who never go wrong. We conclude the paper by explaining how our defence may be able to succeed without assuming (a), in a way that is consistent with compatibilism.

Introduction

Despite the common belief (at least amongst philosophers of religion) that Alvin Plantinga solved the logical version of the problem of evil (LPOE)¹, recent criticism may show Plantinga ’s (1974) free will defence (FWD) to be unsound. Crucial to his defence is the existence of true counterfactuals of creaturely freedom and the possibility of universal transworld depravity (UTD), which have been the target of this criticism.² In this paper, we shall not defend Plantinga’s argument. Instead, we shall argue that there is a simpler and more plausible version of the FWD that establishes that the existence of God is consistent with the existence of evil, given a libertarian conception of freedom. In particular, our argument does not require that there be true counterfactuals of creaturely freedom or the possibility of UTD. The only two assumptions required by the argument are (a) that libertarianism (construed to require the Principle of Alternative Possibilities) is possibly true, and (b) that God’s existence is consistent with the existence of significantly free agents who never go wrong. We conclude the paper by explaining how our defence may be able to succeed without assuming (a), in a way that is consistent with compatibilism.

Before we develop our FWD, a few clarifications are necessary. First, we understand by libertarianism the conjunction of the following two propositions:

• 1. Incompatibilism is true, i.e. causal determinism is incompatible with free action.

• 2. Some people sometimes act freely.

Libertarianism logically entails that determinism is false. According to determinism, a complete, true description of the state of the world at any one time, together with a complete, true description of the laws of nature, completely entails every truth about the world at every later time. Our aim is not to defend libertarianism. We claim only that, given that it is possibly true, the LPOE is defeated. Henceforth, when we talk about free action, we mean it in the libertarian sense that the agent involved is not causally determined to perform the action in question.

We understand by God the perfect being, where a being is perfect only if it has all perfections essentially.³ Something has some property P essentially only if it cannot lack P, i.e. if the state of affairs consisting in that thing’s having P’s complement is an impossible state of affairs. Since omnipotence, omniscience, and moral perfection are perfections, it follows that God has these properties essentially.

Finally, we understand by evil any bad state of affairs, e.g. Paul’s having a minor toothache. The proponent of the LPOE claims that the following two propositions are incompatible:

• 3. God exists.

• 4. There is evil in the world.

Specifically, the proponent of the LPOE claims that these two propositions are logically inconsistent-that a contradiction can be derived from (3) and (4), together with descriptions of the relevant properties or some set of allegedly logically necessary truths. This, together with the obvious fact that there is at least one instance of evil, entails that God does not exist.

Plantinga’s version

One way to show that (3) and (4) are consistent is to identify a proposition $R$ that is consistent with (3) and that entails (4). Plantinga’s candidate proposition for $R$ was,

• 5. God actualizes a world containing moral good and every creaturely essence suffers from transworld depravity.

Let us say that a state of affairs S includes a state of affairs S* if it is impossible that S obtain and S* fail to obtain. Let us further say that an agent is significantly free if she is free with respect to some morally significant actions (e.g. whether to keep a promise). According to Plantinga (1974, p. 188), an essence⁴ E suffers from transworld depravity if and only if, for every world W such that E entails the properties is significantly free in W and always does right in W ⁵, there is a state of affairs T and an action A such that,

• (a) T is the largest state of affairs God strongly actualizes in W,

• (b) A is morally significant for E’s instantiation in W, and

• (c) If God had strongly actualized T, E’s instantiation would have gone wrong with respect to A.

Plantinga claims that (5) is possibly true. It should be evident that if he were right about this, then (3) and (4) would be consistent, for the conjunction of (3) and (5) entails (4). One problem with this compatibility proof is that Plantinga gives us no reason whatever to suppose that (5) is a genuine possibility; he simply asserts that it is. One might argue that since it is not known that (5) is possible, neither is it known from (5) that (3) and (4) are consistent.

Otte (2009) has argued that (5) is necessarily false and so is compatible with no proposition, including (3). Plantinga (2009) agrees that the FWD, as he initially stated it, is flawed in light of this criticism. Otte did, however, suggest revisions of Plantinga’s statement of transworld depravity that are immune to his criticisms. Unfortunately for this FWD, Pruss (2012) offers a counterexample that ‘work(s) just as well against the refined versions of transworld depravity given by Otte.’ If this is right, then the FWD as Plantinga stated it is in trouble, and the LPOE is back on the table as an objection to theism, for we are left without a reason to suppose that (3) and (4) are consistent. In what follows, we shall prove that we need not mention anything about transworld depravity or there being true counterfactuals of creaturely freedom (such as in (c) above) in order to show that free will defeats the LPOE. Our defence is, therefore, immune to these recent objections.

A simpler version

Suppose that we are inclined to think that the following principle is true:

(AP) A person P performs some action A at some time $t$ freely only if it is possible for P to have freely refrained from performing A at $t$.⁶

The following is an example of what AP requires. Let us say that John is free in W. At $t_{1}$, John must decide to perform A or $\sim$A. Thus, there is a possible state of affairs S that obtains in W that (a) includes John’s being faced with a decision to perform A or $\sim$A at $t_{1}$, (b) does not include John’s deciding to perform A, and (c) includes John’s being free with respect to A. Furthermore, suppose John freely performs A at $t_{2}$ in W. By AP, John freely performs A at $t_{2}$ only if there is a possible world W* that (a) includes S, (b) shares W’s past and is identical to W in every way possible prior to $t_{1}$, and (c) includes John’s decision to perform $\sim$A at $t_{2}$. Put otherwise, AP entails that John acts freely on this occasion only if there is at least one distinct possible world that shares an initial world segment with W but that departs at the moment of decision by virtue of John’s freely making the opposite choice.⁷

Now we are ready to see why AP gives us a reason to think the LPOE is unsound. We understand by a saintly agent an agent who (i) is significantly free (i.e., an agent who freely performs some morally significant actions), and (ii) never goes wrong with respect to any of her actions (i.e., who always freely makes the right decision). Consider the following candidate for (5)’s replacement:

• 6. Saintly agents exist.

Is (3) consistent with (6)? Given the truth of AP, it certainly seems so. A morally perfect being might have good reasons for creating significantly free agents. God might, for example, want to create significantly free agents in order that there might be moral good, which we understand as good that comes of a free decision to do what is right when the alternative was possible. This is a good not even God can actualize on his own, for his being essentially morally perfect eliminates the possibility of his going wrong. Perhaps he might want to create people with whom he can have a meaningful and loving relationship, and a requirement of this is that they are significantly free. In any case, it is clearly conceivable that God creates saintly agents. Even if conceivability does not entail possibility, we can still say that it is evidence of possibility. If this is right, then we have evidence from conceivability to suppose that (3) and (6) are consistent.⁸

Indeed, some atheists have thought that the fact that God can create saintly agents supports the LPOE. Mackie (1955), for example, famously asked, ‘if God has made men such that in their free choices they sometimes prefer what is good and sometimes what is evil, why could he not have made men such that they always freely choose the good?’ The presumption was that God can, if he exists, create saintly agents, and that this undermines a defence from free will. Contra Mackie, we now argue that if God can create saintly agents—or any significantly free agents—then the LPOE is unsound.

Suppose that God actualizes a world W in which there is no evil whatever and in which the only agents that God creates are saintly. All of the agents in W are agents who never go wrong with respect to any of their actions. Eve is one of these saints. God orders Eve to eat the fruit from the tree in the garden (perhaps because it would be good for her in some way) and Eve promises to do so. Let A be the action Eve keeps her promise to God. At time $t_{1}$, Eve must decide whether to A or $\sim$A, and A is such that, should Eve decide against performing A, she (i) will go wrong, and (ii) will have disobeyed God. That is, deciding to perform A is morally significant for Eve. Thus, there is a state of affairs S that obtains in W that,

• 7. Includes Eve’s being faced with a decision, at $t_{1}$, to perform A or $\sim$A,

• 8. Does not include Eve’s deciding to perform A,

• 9. Includes A being such that, should Eve decide not to perform A, she (i) will go wrong, and (ii) will have disobeyed God, and

• 10. Includes Eve’s being free with respect to A.

Eve decides to perform A at $t_{2}$, whereupon she performs A at some later time $t_{3}$. By AP, Eve is free with respect to her performing A only if there is some other world, W*, that

• 11. Includes S,

• 12. Is identical to W in every way possible prior to $t_{2}$,

• 13. Includes Eve’s decision to perform $\sim$A at $t_{2}$, and

• 14. Includes Eve’s failure to freely perform A at $t_{3}$.

S includes A being such that, should Eve decide against performing A, she (i) will go wrong and (ii) will have disobeyed God. Consequently, Eve goes wrong at least once in W*. This, of course, entails that there is an instance of evil in W*, for the state of affairs consisting in Eve’s going wrong is a bad state of affairs. Here is the upshot: Eve can disobey God in W* only if God exists (since a person cannot disobey a non-existent being). Moreover, W* shares W’s past prior to $t_{2}$, which includes God’s existence (God’s existence is a hard fact in W prior to $\hbox {t}_{1}$). It follows that W* includes God’s existence and an instance of evil. Hence, the consistent conjunction of (3) and (6) entails that (3) and (4) are compatible if AP is true.

Compare this simple compatibility proof with Plantinga’s FWD. Unlike Plantinga’s version, we do not assume that there are any true counterfactuals of creaturely freedom or that God knows them by virtue of his being omniscient (both of which are controversial). We do not assume that it is possible that every creaturely essence suffers from transworld depravity. The argument that we have sketched avoids all criticisms associated with the above theses. The only two assumptions required by this FWD are that AP is possibly true and that (3) and (6) are consistent, the latter of which is required by Plantinga’s version as well. It is in this sense that the argument that we have produced is a simpler (libertarian) FWD.

Interestingly, we may not even require the first of these two assumptions, that is, the assumption that AP is possibly true. The above defence may work just as well even if compatibilism is true and AP is false. Compatibilism says that free will is compatible with determinism; that there is a possible world in which people sometimes act freely and in which determinism is true. It does not say that free action is incompatible with indeterminism. It, therefore, does not imply the negation of (15):

• 15. Free action is compatible with indeterminism.

Although compatibilism does not entail (15), it is consistent with it. One interesting fact about the above defence, then, is that it can be advanced in a way that is consistent with compatibilism. In other words, neither incompatibilism nor AP need be assumed.

Whether or not compatibilism is true, God can actualize indeterministic worlds in which some people are sometimes faced with morally significant decisions. He can place people in situations in which they have genuinely possible alternative options. Moreover, it is consistent with compatibilism to suppose that, although alternative possibilities are not necessary for free action, they may still produce value in the world. One can maintain that someone’s choosing the good when they really could have refrained from doing so is more impressive than someone’s choosing the good when doing otherwise was not genuinely open to them. If this is right, then, even given compatibilism, God might have reason to create free creatures who sometimes must make morally significant decisions in indeterministic situations.

Here is how the compatibilist-friendly defence would go: God actualizes an indeterministic world and puts Eve in an indeterministic situation in which she must decide whether or not to keep a promise to God. Suppose she in fact decides to keep her promise. Given that the world is indeterministic, and that, therefore, the other option is open to her (the option to break her promise to God), there is another possible world in which Eve fails to keep her promise to God. In that world, God and evil exist. Notice that nothing in this story entails AP, libertarianism, or the falsity of compatibilism. If (15) is true, then all the theist has to do is to identify a reason God might have for creating people who must make morally significant decisions in indeterministic situations.

Given the requisite assumptions, the LPOE is unsound, and we need not appeal to anything like transworld depravity or counterfactuals of creaturely freedom in order to demonstrate this.