Michael J. Almeida offers two criticisms of the argument of my ‘A trilemma for divine command theory’. The first criticism is that I mistakenly assume the validity of the following inference pattern: property A is identical to property B; property B supervenes on property C; therefore, property A supervenes on property C. The second criticism is that I have misinterpreted the moral-supervenience thesis upon which I rely in making this argument. The first of Almeida's criticisms is completely untenable. The second of his criticisms casts doubt on my argument, a doubt that I can mitigate but not entirely dispel.

Following Linda Zagzebski's discussion of the paradoxical implications of moral luck for Christian morality, I explore the role of religious luck in two accounts of divine election – that of Paul the Apostle and that of the sixteenth-century Jewish thinker, Rabbi Judah Loeb of Prague. On both accounts, special religious status is conferred unrelated to the deserts of the beneficiary. What sense does it make to ascribe religious worth to someone if it simply came his way? Both accounts appeal to the notion of religious virtue to answer this question. On Rabbi Judah's account, like ethical virtues, religious virtues inhere within the essence of their bearer and thus belong to him necessarily; on Paul's they are accidental, a matter of luck. Thus Paul's account, more than Rabbi Judah's, suffers from the paradox troubling Zagzebski.

In a number of earlier papers I have attempted to defend the providential utility of simple foreknowledge as a via media between the accounts of divine providence offered by Molinists, on the one hand, and ‘open theists’, on the other. In the current issue of this journal, Michael Robinson argues that my response to one of the standard difficulties for simple foreknowledge – that its providential employment would generate explanatory loops – is inadequate. In the following paper I answer Robinson's charge.

In this article I respond to Thomas Talbott's criticisms of the view of hell I have defended. In particular, I argue that coherent sense can be made of the choice to be eternally separated from God. Moreover, Talbott does not successfully show how God can save everyone without overriding their freedom. Finally, I argue that there is no significant sense in which sinners defeat God or sin with impunity on the view I have defended. Talbott's case that universalism necessarily follows from God's perfect love and power then fails.

In this brief reply to Walls's challenging critique, I try to do two things: first, clarify the most fundamental point on which I think Walls and I disagree, and second, argue that, as surprising as it may first appear, Walls's free-will theodicy of hell requires that God interfere with human freedom in inappropriate ways.

Finite spirits can be plausibly viewed as entities postulated by a theory, comparable to the position on mental states and processes developed in the latter part of the twentieth century. This position is developed here by reference to the account in the synoptic gospels of the exorcism of the Gadarene demoniacs. The role played by specifying causal relationships between postulated entities and objects whose existence is not in doubt is examined. Also, various features of theories are discussed in relation to this example, viz. theory-laden description, classifying theories as naturalistic or supernaturalistic, kinds of evidence, and the importance of the method of hypothesis (or abduction) in critically scrutinizing the claims of religion.

In his reply to my original essay, David Hunt maintains that I do not discuss how his defence of providentially useful simple foreknowledge violates the Metaphysical Principle. Further, he claims that I try to force him into both affirming and denying the accidental necessity of future events and their role in explaining divine advice-giving. In this response, I attempt to articulate more fully why Hunt's defence of simple foreknowledge implies that dependency loops could unfold. Further, I argue that Hunt's scenario is not tenable, whether one affirms that future events are accidentally necessary or contingent.

The celebrated free-will defence was designed to show that the ideal-world thesis presents no challenge to theism. The ideal-world thesis states that, in any world in which God exists, He can actualize a world containing moral good and no moral evil. I consider an intriguing two-stage argument that Michael Bergmann advances for the free-will defence, and show that the argument provides atheologians with no reason to abandon the ideal-world thesis. I show next that the existence of worlds in which every essence is transworld untrustworthy provides atheologians with no better reason to abandon the ideal-world thesis. I conclude that neither the free-will defence nor Bergmann's revised free-will defence is a convincing response to the atheological challenge.

Many people claim to have had direct perceptual awareness of God. William Alston, Richard Swinburne, Gary Gutting, and others have based their philosophical views on these reports. But using analogies from our encounters with humans whose abilities surpass our own, we realize that something essential is missing from these reports. The absence of this element renders it highly unlikely that these people have actually encountered a divine being.

In this brief rejoinder to Talbott's reply, I argue that his clarifications pose a dilemma for him: he must either modify his account of unbearable misery, or give up his claim that all sinners must reach a point where they can resist God no farther.

This paper provides a comment on Brian Zamulinksi's article in Religious Studies, 39 (2003), 43–60. Contrary to Zamulinski's claim that religions are not truth-oriented but function as fictions, it is contended that they could not serve the purpose he assigns them unless their adherents regarded them as true. Religions must therefore be truth-oriented. The substantive question is whether any of them are true, and Zamulinski's paper provides no new method for addressing this question.

In reply to Mawson, I accept that each and every religion includes the self-referential belief that it is true. I seek to show that this admission does not entail that the rest of the beliefs of religions track the truth or that they are not better explained through the religion-as-fiction hypothesis. If that hypothesis is well-grounded, it gives us good reason not to take arguments for religions' non-self-referential beliefs seriously.