1 For some helpful distinctions between various forms of the problem of evil, see Adams, Robert M. ‘Plantinga on the Problem of Evil,’ in Tomberlin, J.E. and van Inwagen, P. eds., Plantinga (Dordrecht: D. Reidel 1985)Google Scholar. Note that Adams calls this form of the problem the ‘logical abstract’ form. I prefer the terminology I have introduced for reasons that will soon become apparent. (Adams has told me that he also prefers my terminology.)

2 The Nature of Necessity, hereafter abbreviated as ‘NN’ (Oxford: Oxford University Press 1974), 165

3 Adams, ‘Plantinga on the Problem of Evil,’ 226. Adams goes on to show why Plantinga’s responses to the other forms of the problem are not as satisfying.

4 God, Freedom, and Evil, hereafter abbreviated as ‘GFE’ (Grand Rapids, MI: Eerdman’s 1974), 11; cf. NN, 164.

5 GFE, 12; NN, 164

6 Mackie, ‘Evil and Omnipotence,’ Mind 64 (1955), 200-1; quoted by Plantinga in GFE, 16

7 See NN, especially 1-9, for Plantinga’ s characterization of ‘broadly logical’ modality. Plantinga equates his notion of ‘broadly logical’ modality with ‘metaphysical’ modality in Does God Have a Nature? (Milwaukee: Marquette University Press 1980), 102.

8 See Plantinga’s discussion of the Ontological Argument in NN, 196-221.

9 In NN, 165, n. 1, Plantinga refers the reader to God and Other Minds (Ithaca: Cornell University Press 1967), ch. 5, for Stage I. GFE contains both Stage I and Stage II.

10 Of course, it would be silly to require that PPP always be defeated by an argument. In some cases (e.g. 2 + 2 = 5, and even much less clear cases), PPP could be defeated by the intuitive plausibility of the impossibility of the proposition itself.

11 I am assuming, as Plantinga does, that all the inferences allowed by 55 are valid inferences when the type of necessity and possibility involved is Plantinga’s ‘broadly logical’ type. If 55 inferences are allowed, then it can be proven that no proposition of the above forms can be contingent. Of course, we do not want simply to say that only contingent propositions can receive the Presumption, because whether a proposition is contingent is precisely what we are trying to decide by applying the Presumption to it. The point I am making is that it is logically impossible (strictly logically impossible, as opposed to Plantinga’s notion of ‘broadly logical’ modalities) that a proposition of one of these forms be contingent. For a discussion of restricting the presumption to those propositions that can be contingently true, see Robert M. Adams, ‘Presumption and the Necessary Existence of God,’ NOUS 22 (1988), 21-2.

12 See Bennett, A Study of Spinoza’s Ethics (Indianapolis: Hackett 1984), 72. Bennett’s restriction is stronger than the preceding restrictions because not all propositions with modal concepts nested within them are such that their possible truth entails their necessary truth. Consider At least one glimph exists, where a ‘glimph’ is defined as a lion which is essentially brown—a lion that is such that it is necessarily true that if it exists, then it is brown. Since ‘glimph’ is a modal concept, Bennett would not allow PPP to apply to At least one glimph exists. But since that proposition’s possible truth does not entail its necessary truth, it would escape the preceding restrictions.

13 This issue will be discussed in Section V, below.

14 NN, 165; cf. GFE, 26.

15 Plantinga mentions some other candidates, but criticisms very similar to those below would apply to these as well.

16 This is thought not to jeopardize God’s omnipotence since it is not possible to cause a creature to freely do what he wouldn’t do.

17 And if it is a fact that all the worlds in which I do good but no evil are unrealizable, this is a contingent fact.

18 Thus these suppositions would also tell against another of Plantinga’s candidates for r, which is the conjunction of (32) and

19 Thus, the possibility of another of Plantinga’s candidates for r, the conjunction of (33) and (34) (see NN, 190), could also be called into question by the supposition that there are necessary relations between facts about what possible creatures would freely do.

20 I don’t.

21 And, in fact, I think that most of what pass for intuitions of metaphysical possibility are really no more than our not being able to see anything threatening the possibility of a proposition. I believe that our intuitive access to modal facts is largely through intuitions of necessity and impossibility.

22 On a ‘broadly logical’ or ‘metaphysical’ reading of ‘could have been,’ where more is required than a mere lack of contradiction in supposing otherwise (see Section II, above). I suppose there is no contradiction involved in any of these combinatorial possibilities, so they are all ‘analytically possible.’

23 Plantinga construes (1) such that this implication holds; see NN, 165.

24 See NN, especially 213-17.

25 NN, 216; my emphasis

26 See NN, 216-21.

27 Of course, we do have a good reason for thinking that (2) is possibly true: it is actually true!

28 More expansively, one is saying that there is possible world such that, if it were actual, then it would have been the case that God was unable to actualize any of the possible worlds which contain moral good but no moral evil.

29 And I do think that Stage I of Plantinga’s case is quite impressive.

30 This paper was first presented at a Fall 1987 seminar at UCLA on the problem of evil. I am grateful to the co-leaders of that seminar, Marilyn McCord Adams, and Robert Merrihew Adams, for their help and advice on this project.