Jerome Gellman has recently disputed my claim that a naturalistic explanation for mystical experiences is available, a better explanation than any current attempt to show that God is sometimes perceived in those experiences. Gellman argues (i) that some mystics do not ‘fit’ the sociological explanation of I. M. Lewis; (ii) that the sociological analysis of tribal mysticism cannot properly be extended to theistic experiences; and (iii) that mystical experiences merit prima facie credence, so the burden of proof falls on the naturalist. I reply (i) that the alleged counter-examples either do fit Lewis's explanation or are too poorly known to judge; (ii) that Lewis's theory, supplemented by recent neurophysiological findings, provides a strong explanation for all mystical experiences; and (iii) that the burden of proof, if there is one, now falls on the theist.

In this paper, I argue that if the libertarian free will defence were seen to fail because determinism were seen to be true, then another solution to the problem of evil would present itself. I start by arguing that one cannot, by consideration of agents' choices between morally indifferent options, reach any conclusion as to these agents' moral qualities. If certain forms of consequentialism were false, determinism true, and if there were a God who chose to create this universe, then His choice would have been between such options. Consideration of the general nature of the universe God putatively chose to create would not then license any conclusion as His moral qualities.

In this response to Stenmark's critique of my views on rational theology, I concentrate on his distinction between the epistemic and the practical goals of religion and between descriptive and normative rational theology. With regard to the first distinction, I grant that truth claims play an essential role in religious belief and that it is indeed the task of philosophy of religion to decide on the meaning and rationality of such claims. I argue, however, that since such claims are internally related to the practical context of religious belief, their meaning and rationality cannot be determined apart from this context as is done in the kind of rational theology which Stenmark calls ‘scientific’. With regard to the second distinction, I reject Stenmark's view that philosophy of religion has a descriptive task with reference to religion, and hence also his claim that I have put forward a false description of ‘the religious language game’.

‘An understanding with the reader’ which appears as an appendix to Kierkegaard's Concluding Unscientific Postscript has puzzled readers by apparently ‘taking back’ everything that the pseudonymous author Climacus has said in the text that precedes it. I agree with John Lippitt that this ‘revocation’ should be read as a humorous statement; however, I disagree with Lippitt's assessment that the tone of the revocation is essentially ‘non-urgent’ and ‘modest’. I propose that a fuller picture of Climacus's character is needed to understand the spirit of the revocation and its relationship to the rest of the text.

In a recent article in this journal Andrew Chignell assesses attempts by Marilyn McCord Adams and Eleonore Stump to resolve the problem that infant suffering poses for theistic belief, concluding that while the theodicy of each is inadequate in its current form, both can be satisfactorily amended. I argue that (1) Chignell fails to show that the theodicy of either Adams or Stump is inadequate and that (2) since Chignell's revisions are based on assumptions about God and evil held by few, such revisions are of little value as responses to the actual challenge infant suffering poses for theistic belief.