There are good reasons to think that at least a part of Hume's project in the ‘The natural history of religion’ was to buttress a philosophical critique of the reasonableness of religious belief undertaken in other works, and to attack a fundamentalist account of the history of religion and the foundations of morality. But there are also problems with supposing that Hume intended to achieve either of these goals. I argue that two problems in particular – accounting for Hume's neglect of revelation, and accounting for his remarks on the ‘invincibility’ of the reasons for ‘genuine theism’ – can only be resolved by recognizing that Hume's purposes in ‘The natural history’ were not fundamentally critical. If I am right, Hume's purpose was mainly to explain why ‘false’ or ‘adulterate’ forms of religious belief are so widespread and so influential.

In this article I assess the coherence of Jonathan Edwards's doctrine of divine simplicity as an instance of an actus purus account of perfect-being theology. Edwards's view is an idiosyncratic version of this doctrine. This is due to a number of factors including his idealism and the Trinitarian context from which he developed his notion of simplicity. These complicating factors lead to a number of serious problems for his account, particularly with respect to the opera extra sunt indivisa principle. I conclude that Edwards sets out an interesting and subtle version of the doctrine, but one which appears mired in difficulties from which he is unable to extract himself.

This is a new argument to the effect that religions are not truth-oriented. In other words, it is not a fundamental function of religion to represent the world accurately. I compare two hypotheses with respect to their likelihood (in A. W. F. Edwards's technical sense). The one which entails that religion is not truth-oriented is a better explanation than its competitor for a number of empirical observations about religion. It is also at least as probable. I point out that, once one has established that religions are not truth-oriented, it is possible to argue that religions are false and it is possible to run a sound ad hominem argument against religious believers who advance religious claims. I suggest that the results are early ones and that what matters is evaluating religion in the way I illustrate in this paper. The ad hominem argument shows that the question of whether religion is truth-oriented is particularly important.

According to the proportionality objection to hell, infinite suffering is out of proportion to any wrong that finite human beings could commit and is hence unjust and inconsistent with God's moral perfection. The continuing-sin response concedes that eternal consignment to hell is out of proportion to the sins people commit during their earthly lives, but argues that people in hell continue to sin while in hell and, in this way, extend their consignment to hell ad infinitum. In this essay, I evaluate the continuing-sin response. In particular, I argue that whether there is a proportionality problem to begin with and whether the continuing-sin response succeeds as a response depends on the character of the suffering that is experienced in hell.

Descartes holds that God's will is immutable. It cannot be changed by God and, because He is supremely independent, it cannot be changed by anything else. Descartes' God acts by a single immutable will for all eternity, and there is no sense in which it is possible for Him to will or to have willed anything other than what He in fact wills. Passages in which Descartes might appear to be suggesting a different view are simply manifestations of his analytic method.

The Molinist doctrine that God has middle knowledge requires that God knows the truth-values of counterfactuals of freedom, propositions about what free agents would do in hypothetical circumstances. A well-known objection to middle knowledge, the grounding objection, contends that counterfactuals of freedom have no truth-value because there is no fact to the matter as to what an agent with libertarian freedom would do in counterfactual circumstances. Molinists, however, have offered responses to the grounding objection that they believe are adequate for maintaining the coherence of middle knowledge. I argue that these responses to the grounding objection are not adequate, and that what I call the ‘generic grounding objection’ still poses a serious challenge to middle knowledge.

I suggest that Tim Bayne's use of the term ‘inclusion’ to describe the model of the Incarnation found in Morris and Swinburne may have misled him. The experiences of the Word do not include those of Jesus in the way that mine include my experiences as a teenager; but He is aware, in the case of Jesus, that ‘these experiences are mine’, which is not true of His awareness of the experiences of other people. Again, Bayne rejects the idea that what differentiates the experiences of Jesus from those of the Word is that they differ in kind, on the grounds that they are integrated so as to be co-conscious in the divine consciousness; but this is only true if we think in terms of ‘inclusion’. Nor are any false beliefs held by Jesus part of the beliefs of the Word. Furthermore (although this is not related to ‘inclusion’) while a single soul may be sufficient to unite experiences, it need not be necessary; some other factor may (and I think does) unite the human and divine experiences of Christ.

I make three points in response to Richard Sturch's comments on my paper: I defend my interpretation of the Morris–Swinburne (M–S) account of the Incarnation; I argue that the M–S model appears to undercut the view that the unity of consciousness can be explained in terms of the self; and third, I argue that M–S model seems to entail that God has false beliefs.