Skip to main content Accessibility help

Probabilistic Arguments from Evil

  • Paul Draper (a1)


Many philosophers have held that traditional theism can be conclusively disproved because it can be shown to be logically inconsistent with some known fact about evil. By ‘traditional theism’ I mean the statement that G. There exists an omnipotent, omniscient and morally perfect person who created the universe.



Hide All

1 See, for example: Hume, David, Dialogues Concerning Natural Religion, ed. Smith, Norman Kemp (New York: Macmillan Publishing Co., 1947), Part X; and Mackie, J. L., ‘Evil and Omnipotence’, Mind, LXIV (1955), 200–12.

2 Pike, Nelson, ‘Hume on Evil’, Philosophical Review, LXXII (1963), 180–97.

3 Plantinga, Alvin, ‘The Probabilistic Argument from Evil’, Philosophical Studies, XXXV (1979), 153.

4 For example, an argument from evil might attempt to show that some known fact about evil makes traditional theism less probable (all things considered) than it would otherwise be. If we let B stand for one's background information, and employ the idealization that epistemic probability can be represented as probability conditional on a statement representing one's total evidence, then the above claim can be represented as follows: P(G/En & B) < P(G/B). This claim does not entail nor is it entailed by the claim that P(G/En) < 1/2.

5 This involves ignoring his interesting suggestion that probabilistic arguments from evil fail because belief in God can be included in the foundations of a rational belief system. I have criticized this suggestion at length in ‘Evil and the Proper Basicality of Belief in God’, Faith and Philosophy, VIII (1991), 135–47.

6 Plantinga, p. 49.

7 Plantinga, p. 49.

8 ‘Reason and Belief in God’. In Alvin Plantinga and Nicholas Wolterstorff eds., Faith and Rationality: Reason and Belief in God. (Notre Dame: University of Notre Dame Press, 1983), p. 22.

9 Some of the theorems of the probability calculus presuppose logical omniscience, and hence are not true for epistemic probability relative to the epistemic situations of human beings. However, since all of the entailment relations that are asserted in this argument (and the rest of this paper) are ones that humans do have knowledge of, the assumption of logical omniscience may be regarded in this context as a harmless idealization. In contrast, Plantinga points out that it follows from the probability calculus that, to show that P(G/E) < 1/2, one would need to show that G is not a necessary truth (Plantinga, ‘The Probabilistic Argument from Evil’, pp. 4–5). If this were offered as a reason for doubting that a successful probabilistic argument from evil could be constructed, then it would be a very weak reason indeed. For here the idealization in question is obviously not harmless – it leads to a plainly false claim.

10 Martin, Michael, ‘Is Evil Evidence Against the Existence of God?Mind, LXXXVII (1978), 429–32.

11 Martin, p. 431.

12 ‘Self-Profile’. In James E. Tomberlin and Peter van Inwagen, eds., Alvin Plantinga (Dordrecht: D. Reidel, 1985), p. 35.

13 ‘Evil and Theodicy’, Philosophical Topics, XVI (1988), 119–32. Earlier versions of Rowe's argument can be found in Philosophy of Religion: An Introduction (Belmont: Wadsworth, 1978), Ch. 6; ‘The Problem of Evil and Some Varieties of Atheism’, American Philosophical Quarterly, XVI (1979), 335–41; and ‘The Empirical Argument from Evil’. In Robert Audi and William J. Wainwright, eds., Rationality, Religious Belief and Moral Commitment (Ithaca: Cornell University Press, 1986), pp. 227–47.

14 In ‘The Inscrutable Evil Defense Against the Inductive Argument from Evil’ (unpublished), James F. Sennett makes a claim very similar to my claim that Rowe's inference is justified only if S is true. However, he defends his claim in a very different way than I defend mine.

15 Rowe, ‘Evil and Theodicy’, pp. 123–4.

16 Hume, pp. 211–12.

17 Hume, p. 205.

18 Although I was not trying to meet Plantinga's challenge at the time, I implicitly employed Hume's strategy in ‘Pain and Pleasure: An Evidential Problem for Theists’, Nous, XXIII (1989), 331–50.

Recommend this journal

Email your librarian or administrator to recommend adding this journal to your organisation's collection.

Religious Studies
  • ISSN: 0034-4125
  • EISSN: 1469-901X
  • URL: /core/journals/religious-studies
Please enter your name
Please enter a valid email address
Who would you like to send this to? *


Full text views

Total number of HTML views: 0
Total number of PDF views: 0 *
Loading metrics...

Abstract views

Total abstract views: 0 *
Loading metrics...

* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.

Usage data cannot currently be displayed