Plantinga's Free Will Defence: Critical Note

A critical review of Alvin Plantinga's Free Will Defence against the logical Problem of Evil

Citation Information

Allan, Leslie 2015. Plantinga's Free Will Defence: Critical Note, URL = <>.

Alvin Plantinga at Mayo Clinic in Rochester, Minnesota.

In his book, God, Freedom and Evil, Alvin Plantinga argues that the existence of evil is logically compatible with the existence of God. Plantinga defines God as an omnipotent, omniscient and omnibenevolent being. His argument is not a theodicy, aiming to explain why God in fact allows evil. His is a weaker defense, aiming to show that God's existence is logically compatible with some evil [1975: 28]. As such, Plantinga's defense does not serve to explain the actual amount and distribution of evil we see in the real world. A second point of introduction is that this defense does not seek to explain natural evils; that is, evils wrought by nature. These evils include volcanic eruptions, earthquakes, tornadoes, and so on. In his defense, Plantinga focuses on moral evils only; the evils brought about by moral agents acting freely.

In Chapter 5, Plantinga begins by presenting an argument that attempts to show that it is not within God's power to create any possible world involving free agents. Free will, for Plantinga, is meant in the contra-causal sense of there being 'no antecedent conditions and/or causal laws' determining whether an agent freely performs an act or an omission [1975: 29].

Here, he uses the example of Paul turning down an offer of $500 to buy his aardvark after returning from a hunting expedition. Plantinga asks us to consider an alternative history in which Paul was offered $700 for his aardvark (state of affairs S′) [1975: 40f]. Plantinga asks us to consider two conditionals:

(23) If the state of affairs S′ had obtained, Paul would have accepted the offer.

(24) If the state of affairs S′ had obtained, Paul would not have accepted the offer.

Plantinga's argument crucially depends on one of (23) or (24) being true and the other being false. This must be so, so that Plantinga can argue that the possible world in which the consequent of the false conditional is true necessarily cannot be instantiated by God. To illustrate, Plantinga argues that if (23) is true, God necessarily could not have created a world in which S′ obtains and Paul does not accept the offer.

If it turns out that (23) and (24) do not have opposite truth conditions (one true and the other false), given Plantinga's other assumptions, then his case collapses. My reasons for thinking that (23) and (24) do not have opposite truth values are as follows. In using counterfactual conditionals and ascribing truth values to them, we assume the consequent follows (or does not follow) because of the law-like nature of the situation described in the antecedent. If there were not some regularity in the relations between the states described in the antecedent and the outcome described in the consequent, how would we ever know that a counterfactual was true or false? The if . . . then in the conditional ascribes some kind of necessity to the event described in the consequent.

This necessity could be a physical necessity, based on physical laws. 'If it had rained yesterday, my plants would still be alive' is an example. Many instances of indicative conditionals in ordinary use also follow this causal pattern of implication. (E.g. 'If the refrigerator breaks down, our food will spoil.') On the other hand, the necessity could be of a psychological type, based on psychological laws. Plantinga's example is of just this kind. We would know either (23) or (24) to be true based on our previous knowledge of Paul's dispositions and our understanding of the way human beings interact in a variety of situations. It is only because human beings behave in law-like ways that we can predict how a particular person will choose in situations that do not in fact obtain. In terms of possible world semantics, we could say that a counterfactual conditional describes a possible world just like our actual world, in which the same physical and psychological laws operate, but in which the initial conditions are different to the extent described in the antecedent.

This presumption of a causal relationship between the states described in the antecedent and those given in the consequent is not a discretionary aspect of these kinds of conditionals. This assumption is built into the meaning of subjunctive conditionals. I think we can best appreciate this by comparing the above analysis with the assumption that causation is inessential to understanding counterfactuals.

Book cover: God, Freedom, and Evil by Alvin Plantinga

To help us make this comparison, I will now bring in Plantinga's assumption that human agents have free will in his contra-causal sense. By doing this, we will see what impact the above analysis has on the veracity of Plantinga's Free Will Defense. Let us return to Plantinga's possible worlds example. For Plantinga, under the state of affairs S′, Paul is free to accept the offer or to deny the offer. His choice, for Plantinga, is indeterminate. According to Plantinga, there are no physical or psychological antecedent conditions and causal laws that bring about Paul accepting or not accepting the offer. If this is so, then how are we to understand the implication given in the counterfactual conditionals (23) and (24)? It can't be a material implication, as that would render all counterfactuals true. It can't be a logical implication, as Plantinga himself agrees [1975: 41]. Plantinga has already ruled out a contingent, causal implication. Then what kind of implication is it?

His possible world semantics sheds no real light on this question. Plantinga offers the following brief analysis of a counterfactual conditional. For him, this sentence form specifies a particular outcome 'under a specific set of conditions' [1975: 40]. In Plantinga's examples, each counterfactual specifies whether Paul accepts or does not accept the offer under the conditions that the amount offered is $700, that the going rate for that kind of sale is $650, and so on. Plantinga also includes here the condition that Paul is free in making his decision. This is where I think Plantinga's fundamental error lies. In asserting the truth or falsity of a counterfactual conditional, as I argued above, is to assume that there are causal laws operating that make the outcome necessary.

In addition to the semantic problem, Plantinga is also burdened with an epistemic problem. How would anyone know which of (23) or (24) is true and which is false? What evidence could be adduced for thinking that either is true? These are not just problems for Plantinga's counterfactual propositions. These are problems for understanding any subjunctive conditional in which free agents in the contra-causal sense appear. Given these serious difficulties with Plantinga's analysis on the one hand and the explanatory power of the type of causal analysis of counterfactual conditionals advocated above on the other, it seems prudent to accept that law-likeness is an essential feature of subjunctive conditionals.

To further illustrate the untenability of Plantinga's position, consider another example of an indeterministic system. Instead of thinking about free moral agents, let's use a completely physical system. Imagine a machine that can throw up a coin either to the left or to the right of the machine. Imagine further that the side that the coin lands on after each throw is not determined by the initial state of the machine and the physical laws governing such machines. It could be, for example, that the side the coin lands on is the result of a random quantum fluctuation in one part of the machine. Imagine also that the coin was thrown to the left at a specific time in the past, t1, and landed on 'heads'.

Consider now two counterfactuals:

(1) If at time t1, the coin had been thrown to the right, it would have landed on 'heads'.

(2) If at time t1, the coin had been thrown to the right, it would have landed on 'tails'.

It seems clear that we would have no reason to think either of these counterfactuals is true. Furthermore, it seems pointless even positing one or the other, for there is no necessary connection between the antecedent (the throwing of the coin) and the consequent (the side the coin lands on). The if . . . then in the conditional seems misplaced. But the indeterminism of this hypothetical physical system is no more or less contra-causal than Plantinga's free moral agents.

How are we then to interpret subjunctive conditionals that have as their subject free agents not amenable to causal laws? We could render them as meaningless. An old-fashioned positivist could point out that no empirical evidence can in-principle be adduced to increase the likelihood that either (23) or (24) is true. However, I don't think this conclusion is warranted. My favoured approach is to say that both (23) and (24) are false. And this is for the reason that both are based on a mistaken assumption that the implication is indeterministic. Plantinga's error could be regarded a category mistake.

The upshot here is that whether we take (23) and (24) as having no truth value or as both being false, Plantinga's Free Will Defense does not get off the ground. His initial argument that God could not actualize at least one possible world in which Paul chooses differently crucially depends on (23) or (24) being true [1975: 41f]. Plantinga's following argument using a subjunctive conditional referring to a future event [1975: 42–4] labours under the same confusion. His next example of a morally significant choice involving bribery, continuing through to his discussion of transworld depravity [1975: 45–53] repeats the same error.

My argument in this critical note has been that Plantinga has failed to demonstrate the falsity of the non-theists' contention that God could have created a world containing free agents, albeit with no moral evil. Plantinga's stipulation that only choices made by agents possessing free will of the contra-causal kind are morally good created debilitating problems for his argument. I endeavoured to show that his analysis of counterfactual conditionals is mistaken and that a correct analysis renders the counterfactual propositions in his argument false. As his Free Will Defense requires some counterfactual conditionals about free agents to be true, I argued that his defense does not get to first base.


  • Allan, Leslie 2015. The Problem of Evil, URL = <>.
  • Beebe, James R. 2015. Logical Problem of Evil, Internet Encyclopedia of Philosophy, URL = <> (Retrieved: July 20, 2015).
  • Bradley, R. D. 1967. A Proof of Atheism, Sophia 6/1: 35–49.
  • Plantinga, Alvin 1975. God, Freedom, and Evil, London: Allen and Unwin.
  • Reichenbach, Bruce R. 1981. The Deductive Argument from Evil, Sophia 20/1: 25–42.
  • Tooley, Michael 2015. The Problem of Evil, The Stanford Encyclopedia of Philosophy (Summer 2015 Edition), ed. Edward N. Zalta, URL = <>.

Copyright © 2015

You will be interested in
Is Atheism Dead? by Eric Metaxas
Book cover: Towards the Light: The Story of the Struggles for Liberty and Rights That Made the Modern West by A. C. Grayling
Agnosticism: A Very Short Introduction by Robin LePoidevin
The God Argument: The Case against Religion and for Humanism by A. C. Grayling
A New History of Early Christianity by Charles Freeman
Book cover: Saint Anselm: Basic Writings by trans. S. N. Deane

Share This

  • twitter
  • facebook
  • linkedin
  • googleplus
  • gmail
  • delicious
  • reddit
  • digg
  • newsvine
  • posterous
  • friendfeed
  • googlebookmarks
  • yahoobookmarks
  • yahoobuzz
  • orkut
  • stumbleupon
  • diigo
  • mixx
  • technorati
  • netvibes
  • myspace
  • slashdot
  • blogger
  • tumblr
  • email
Short URL: