A Defence Is Not a Theodicy: Plantinga and Hick Answer Different Questions About Evil — Epoche C1
Alvin Plantinga's free-will defence and John Hick's soul-making theodicy are the two arguments most often produced when a philosopher of religion is asked what religious thought says about evil, and they are routinely read as rival versions of a single reply. They are not. They answer different questions, carry different burdens of proof, and break down at different points. This essay sets out the machinery of each in enough detail to show where each holds and where each gives out, and argues that running the two together is what has made the standard theistic reply appear far weaker than it is. The distinction that organises everything Plantinga himself drew the line, and it is a distinction of type rather than of degree. The two labels are terms of art in this literature and mean the following. A defence answers the question is belief in God consistent with the existence of evil? To succeed it must produce a story that is merely possible and which, if true, would explain why God permits evil. Nobody need believe the story, its author included. A theodicy answers the question why does God in fact permit evil? To succeed it must produce a reason that is plausibly true. That is a far heavier burden, and a theodicy that is merely possible has failed. The difference matters because of the logical shape of the charge each replies to. To show that two propositions are consistent — that they can both be true — it suffices to exhibit a third proposition $r$ such that its conjunction with the first is possible and, together, they entail the second: $$\text{if } \Diamond(p \wedge r) \text{ and } (p \wedge r) \models q, \text{ then } \Diamond(p \wedge q).$$ The reasoning is immediate: in any possible world where $p \wedge r$ holds, $q$ holds too, so $p \wedge q$ holds somewhere, and a pair true together somewhere is not a contradictory pair. Crucially, $r$ need not be true or even likely; it need only be possible. So if the charge against theism is that it is self-contradictory, a single coherent possibility rebuts it completely. If the charge is instead that evil makes theism improbable, possibility is worthless, since improbable things are possible. The two replies below are calibrated to those two different charges. What Mackie actually argued The target of the free-will defence is a specific piece of reasoning, more careful than the summary usually given. J. L. Mackie's 1955 paper in Mind is often described as claiming that three propositions — God is omnipotent, God is wholly good, evil exists — form an inconsistent set. Mackie explicitly denied that. No contradiction is visible among those three; the contradiction appears, he argued, once one adds what he called quasi-logical rules connecting the terms. The two he supplied are: that good is opposed to evil in such a way that a good thing always eliminates evil so far as it can, and that there are no limits to what an omnipotent thing can do. Grant those and the set does become inconsistent, since a being able to eliminate all evil and disposed to eliminate all it can would leave none. Both added premises are where the later literature does its work: Plantinga attacks the second, and the sceptical replies discussed below attack the first. Mackie also anticipated the appeal to freedom, with an objection that is genuinely sharp. If it is logically possible for a free person to choose rightly on one occasion, it is logically possible for that person to choose rightly on every occasion; a world of free beings who always freely do good is therefore a possible world; an omnipotent being can bring about any possible world; so God could have made free beings who never sin. Freedom, he concluded, cannot be what explains evil. Type one: the free-will defence Plantinga's reply, set out most rigorously in The Nature of Necessity (1974), denies the third step of that argument — that an omnipotent being can bring about any possible world — and it does so with a piece of machinery that has to be assembled before the denial makes sense. Start with three terms. A possible world , in this tradition, is a maximal way things could have been: a state of affairs so complete that for every proposition it settles whether it obtains. A person is significantly free with respect to an action if free to perform it or refrain and the action is morally weighty. And a counterfactual of creaturely freedom is a proposition of the form if S were placed in circumstances C, S would freely do A . Such propositions, if they have truth-values, are not up to God: were God to make them true by decree, the actions they describe would be caused rather than free. Their truth independent of God's will is the doctrine of middle knowledge, developed by the sixteenth-century Jesuit Luis de Molina — knowledge lying between God's knowledge of necessities and of what he himself decides. Now the pivotal distinction. God strongly actualises a state of affairs by causing it directly, and weakly actualises one by strongly actualising something else which, given the counterfactuals of freedom that happen to hold, results in it. A free choice cannot be strongly actualised, by definition; the most God can do is arrange the circumstances and let the counterfactual decide the outcome. Omnipotence is thereby limited not by any external power but by what "bringing about" can mean for a free act. Mackie's third step assumed that every possible world is available for God to actualise; Plantinga's point is that the available — feasible — worlds are a subset fixed by counterfactuals God does not control. Against that background, transworld depravity can be defined. A creature suffers from it if, in every set of circumstances in which God could have placed it significantly free, it would have gone wrong at least once. One refinement is needed here: Plantinga's official version attaches the property not to persons but to essences , the individual natures God chooses among when deciding whom to create — becaus