¹ The following is a sampling of recent sources for the theoretical debate. For defences of middle knowledge, see: Plantinga, Alvin, The Nature of Necessity (Oxford: Clarendon Press, 1974), ch. 9Google Scholar; Plantinga, , ‘Replies to My Critics’, in Alvin Plantinga, Profiles: An International Series on Contemporary Philosophers and Logicians, vol. 5, ed. Tomberlin, James E. and Inwagen, Peter van (Dordrecht, Netherlands: D. Reidel, 1985)Google Scholar; Kvanvig, Jonathan, The Possibility of an All-Knowing God, Library of Philosophy and Religion (Houndmills, UK: Macmillan Press, 1986), ch. 4CrossRefGoogle Scholar; Craig, William Lane, The Only Wise God (Grand Rapids: Baker Book House, 1987), ch. 12Google Scholar; Otte, Richard, ‘A Defense of Middle Knowledge’, International Journal for Philosophy of Religion, xxii (1987), 161–69CrossRefGoogle Scholar; Freddoso, Alfred, ‘Introduction’, in Luis de Molina, On Divine Foreknowledge (Part IV of the Concordia), translated with introduction and notes, Freddoso, Alfred J. (Ithaca: Cornell U. Press, 1988)Google Scholar; and Wierenga, Edward R., The Nature of God, Cornell Studies in the Philosophy of Religion (Ithaca & London: Cornell U. Press, 1989), ch. 5.Google Scholar For critiques of middle knowledge, see: Adams, Robert Merrihew, ‘Middle Knowledge and the Problem of Evil’, American Philosophical Quarterly, xiv (04 1977), 109–17Google Scholar; Kenny, Anthony, The God of the Philosophers (Oxford: Clarendon Press, 1979), pp. 61–71Google Scholar; Adams, , ‘The Problem of Evil’, in Alvin Plantinga, op. cit.Google Scholar; Hacker, William, ‘A Refutation of Middle Knowledge’, Nous, xx (1986), 545–57Google Scholar, expanded in God, Time, and Knowledge (Ithaca: Cornell University Press, 1989), ch. 2Google Scholar; and Hunt, David P., ‘Middle Knowledge: the “Foreknowledge Defense”’, International Journal for Philosophy of Religion, xxviii (1990), 1–24.CrossRefGoogle Scholar

² Contributions to the practical debate may be found in a number of the sources cited in note 1; in addition, the following articles by David Basinger should be noted: ‘Human Freedom and Divine Providence: Some New Thoughts on an Old Problem’, Religious Studies, xv (1979), 491–510Google Scholar; ‘Divine Omniscience and Human Freedom: a “Middle Knowledge” Perspective’, Faith & Philosophy, 1 (1984), 291–302CrossRefGoogle Scholar; and ‘Middle Knowledge and Classical Christian Thought’, Religious Studies, xxii (1986), 407–22.Google Scholar A reply to the last of these articles may be found in Gordon, David & Sadowsky, James, ‘Does Theism Need Middle Knowledge?’, Religious Studies, xxv (1989), 75–87.CrossRefGoogle Scholar

³ See Hasker, , ‘A Refutation of Middle Knowledge’, op. cit.Google Scholar; also Basinger, David, ‘Middle Knowledge and Human Freedom: Some Clarifications’, Faith and Philosophy, iv (1987), 330–6CrossRefGoogle Scholar, and Hasker, William, ‘Reply to Basinger on Power Entailment’, Faith and Philosophy, v (1988), 87–90.CrossRefGoogle Scholar

⁴ See Adams, , Kenny, and Hunt, , op. cit.Google Scholar

⁵ See especially Basinger, [1984]Google Scholar, Basinger, [1986]Google Scholar, Kvanvig, , Craig, and Freddoso, , op. cit.Google Scholar; this is also the use for which Molina himself introduced the notion of middle knowledge.

⁶ This point is discussed in some detail in Hunt, , op. cit.Google Scholar

⁷ Skeptics include Adams and Kenny, who have argued that current semantics for subjunctive conditionals make the truth-conditions for such conditionals dependent on the prior specification of an actual world (thus putting them on a par with foreknowledge); Hasker, who has denied that the full providential control that middle knowledge makes possible is theologically necessary or even desirable (Adams also appears to find a ‘venturesome’ God theologically acceptable); and Gordon & Sadowsky, who claim that middle knowledge is not even providentially superior to present knowledge(!).

⁸ Op. cit. p. 127.Google Scholar

⁹ Ibid. p. 138.

¹⁰ ‘“No Other Name”: a Middle Knowledge Perspective on the Exclusivity of Salvation through Christ’, Faith and Philosophy, vi (1989), 172–88.Google Scholar

¹¹ On the suitability of eternal suffering as the consequence of unbelief, he has this to say (p. 176): ‘I do not see that the very notion of hell is incompatible with a just and loving God…. Those who make a well-informed and free decision to reject Christ are self-condemned, since they repudiate God's unique sacrifice for sin. By spurning God's prevenient grace and the solicitation of His Spirit, they shut out God's mercy and seal their own destiny. They, therefore, and not God, are responsible for their condemnation, and God deeply mourns their loss.’ And on the suitability of God proceeding with creation in the face of these consequences, he offers this (p. 185): ‘The happiness of the saved should not be precluded by the admittedly tragic circumstance that their salvation has as its concomitant the damnation of many others, for the fate of the damned is the result of their own free choice.’

¹² Craig, reviews some of the evidence for this claim in ‘No Other Name’, 172–4.Google Scholar

¹³ This is essentially the position worked out by Craig, on p. 184Google Scholar of ‘No Other Name’.

¹⁴ Its fullest development is in The Nature of Necessity, op. cit.

¹⁵ Theory X is the elusive grail of part 4 of Parfit's Reasons and Persons (Oxford: Clarendon Press, 1984).Google Scholar Like Parfit, God would presumably reject ‘the Average Principle’, and so would not assign first place to a world containing just one person who is immensely saintly; see ‘No Other Name’, 182–3.Google Scholar

¹⁶ Roughly, a world W is feasible for an agent A iff there is something A can do such that, were A to do it, W would be actual (but wouldn't be actual otherwise).

¹⁷ ‘No Other Name’, 184.Google Scholar

¹⁸ Of course, the ‘contingent damnation’ referred to here should not be confused with logically contingent damnation, since even the transworldly damned (with whom I am contrasting the contingently damned) are such that there is some logically possible world in which they freely accept Christ and escape damnation. Like transworld damnation, contingent damnation is to be defined in terms of feasible (rather than logically possible) worlds.

¹⁹ ‘No Other Name’, 184.Google Scholar This is Craig's proposition (9).

²⁰ Ibid. p. 187.

²¹ Ibid. p. 186.

²² Ibid. p. 184; see also p. 181.

²³ Ibid.

²⁴ This argument, of course, does not demonstrate that there can be no good reason for undertaking evangelical endeavours if Craig's position is accepted. For it might be the case that I receive great personal satisfaction from sharing the gospel with others; it might also be the case that, while someone else would have shared the gospel with Jill if I had not done so, my taking the initiative in the matter freed this person to perform other worthwhile actions instead. But clearly none of this is relevant to Craig's thesis, which requires that my evangelical efforts might make a difference to someone's salvation – i.e. that it is possible for these efforts to result in someone being saved who would not otherwise have been saved. What the argument shows is that this sort of difference will not be made, regardless of any expected differences in my personal satisfaction, or in some third party's achievement of some nonevangelical good.

²⁵ ‘No Other Name’, 177.Google Scholar

²⁶ Hick, John, Evil and the God of Love, rev. ed. (San Francisco: Harper & Row, 1978), p. 255.Google Scholar

²⁷ Op. cit. p. 138.Google Scholar

²⁸ For a recent treatment of the problem of evil which relies heavily on appeals to God's nondeceptive nature, see Inwagen, Peter van, ‘The Magnitude, Duration, and Distribution of Evil: a Theodicy’, Philosophical Topics, xvi (1988), 161–87.CrossRefGoogle Scholar In each case where this appeal is made, however (God's failure to restore us to our original state before the Fall (p. 172), to protect us from the harmful consequences of our bad choices (pp. 173, 177), and to preserve us from natural disasters (p. 181)), van Inwagen cites additional reasons for God's forebearance that have nothing to do with nondeceptiveness, but instead draw attention to how such intervention might be harmful in the long run for the agents concerned.

²⁹ It is also of some (slight) interest that one of the major ‘proof texts’ for middle knowledge cited by Molina (and also by Craig) appears to conflict with Craig's soteriological proposal. At Matthew 11:21–4 Jesus is represented as saying that certain miracles he had performed in Korazin, Bethsaida and Capernaum were such that, had they been performed in Tyre and Sidon, ‘they would have repented long ago in sackcloth and ashes’, and had they been performed in Sodom, ‘it would have remained to this day’. The passage in its entirety certainly suggests that at least some residents of Tyre, Sidon and Sodom, though actually damned, were not transworldly damned. See Molina, , op. cit., p. 116Google Scholar (Disputation 49, sec. 9), and Craig, , The Only Wise God, p. 132.Google Scholar

³⁰ For another recent defence of exclusivism, also sympathetic to Molinism, see Walls, Jerry L., ‘Is Molinism as Bad as Calvinism?’, Faith and Philosophy, vii (1990), 85–98.CrossRefGoogle Scholar Walls argues that the actual world, despite apparent inequities in life-circumstances, might nevertheless include an adequate chance for all, since God could provide an ‘optimal measure of grace’ at a point Walls identifies variously as ‘at the moment of death’, ‘after death’, and ‘during the passage of death’ (p. 93). Thus he seems to endorse some combination of my options (i) and (ii).

³¹ See Hunt, , op. cit.Google Scholar, for a discussion of the grounding issue in light of the Molinists' commitment to libertarian freedom.

³² Otte, , op. cit. p. 167.Google Scholar See also Kvanvig, , op. cit. pp. 136–7Google Scholar, and Craig, , The Only Wise God, p. 140.Google Scholar The best response to the grounding objection, however, is to be found in Wierenga, , op. cit.Google Scholar; I hope to address its arguments on another occasion.

³³ Alvin Plantinga, p. 579.Google Scholar

³⁴ For a recent defence of a risk-taking God, see Hasker, , God, Time, and Knowledge, ch. 10.Google Scholar

³⁵ I wish to thank Bill Hasker for his helpful comments on an earlier draft of this paper. His own reply to Craig, ‘Middle Knowledge and the Damnation of the Heathen’, which focuses on the evangelical implications of Craig's position, is forthcoming in Faith and Philosophy.